WO2023016539A1 - Network path processing method and apparatus, storage medium, and electronic device - Google Patents

Network path processing method and apparatus, storage medium, and electronic device Download PDF

Info

Publication number
WO2023016539A1
WO2023016539A1 PCT/CN2022/112000 CN2022112000W WO2023016539A1 WO 2023016539 A1 WO2023016539 A1 WO 2023016539A1 CN 2022112000 W CN2022112000 W CN 2022112000W WO 2023016539 A1 WO2023016539 A1 WO 2023016539A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
path
network
original
traffic
Prior art date
Application number
PCT/CN2022/112000
Other languages
French (fr)
Chinese (zh)
Inventor
吕田田
吴艳芹
张乐
Original Assignee
中国电信股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国电信股份有限公司 filed Critical 中国电信股份有限公司
Publication of WO2023016539A1 publication Critical patent/WO2023016539A1/en

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/12—Shortest path evaluation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/24—Multipath

Definitions

  • the present disclosure relates to the technical field of network communication, and in particular, to a network path processing method and device, a storage medium, and electronic equipment.
  • the service flow usually selects a round-trip path during operation. If the circuit quality is poor or interrupted, the line will be reselected.
  • SLA Service Level Agreement, Service Level Agreement
  • 5G Fifth Generation Mobile Communication Technology
  • the innovative iFIT The in-situ Flow Information Telemetry (Flow Information Telemetry) scheme came into being. This solution will analyze the poor or interrupted service quality based on the end-to-end follow-the-flow detection mode, and then switch to the hop-by-hop detection mode to locate circuit problems and re-plan routes.
  • a method for processing a network path including: acquiring original traffic data of the original network path and candidate traffic data of the candidate network path, and processing the original traffic data and the The alternative traffic data is subjected to traffic superposition processing to obtain superimposed traffic data; the network path data of the candidate network path is obtained, and path evaluation processing is performed on the superimposed traffic data and the network path data to obtain path evaluation data; network preference data corresponding to the path evaluation data, and perform path recommendation processing on the path evaluation data and the network preference data, so as to determine the optimal candidate path of the original network path among the candidate network paths.
  • the obtaining the original flow data of the original network path and the candidate flow data of the candidate network path includes: acquiring the original path data of the original network path and the candidate path data of the candidate network path, and Perform data cleaning processing on the original path data to obtain original cleaning data; perform data feature calculation on the original cleaning data to obtain original data features, and compare the original data features with the original path data to obtain a data comparison result; if the The result of the data comparison is that the characteristics of the original data do not meet the requirements of the service level agreement SLA performance evaluation data included in the original path data, determine that the original network path is abnormal, and obtain the original traffic data in the original path data and alternative flow data in the alternative path data.
  • the raw data characteristics include at least one of one-way delay, two-way delay, packet loss rate, and bandwidth utilization rate.
  • the SLA performance evaluation data when the original data feature includes the one-way delay, the SLA performance evaluation data includes the maximum one-way delay; when the original data feature includes the two-way delay In this case, the SLA performance evaluation data includes the maximum two-way delay; when the original data feature includes the packet loss rate, the SLA performance evaluation data includes the maximum packet loss rate; when the original data feature includes the packet loss rate In the case of the above-mentioned bandwidth utilization ratio, the SLA performance evaluation data includes the maximum bandwidth utilization ratio.
  • performing traffic superposition processing on the original traffic data and the candidate traffic data to obtain the superimposed traffic data includes: The alternative type data is compared to the message type to obtain the type comparison result; if the type comparison result is that the original type data is the same as the alternative type data, the original traffic data and the alternative traffic data are performed Flow superposition processing to obtain superimposed traffic data; if the type comparison result is that the original type data is different from the alternative type data, superimposing and updating the original path data and the alternative path data to obtain superimposed traffic data .
  • performing traffic superposition processing on the original traffic data and the alternative traffic data to obtain superimposed traffic data includes: performing traffic on the upstream traffic of the original traffic data and the upstream traffic of the alternative traffic data superposition processing, to obtain the upstream traffic after traffic superimposition; perform traffic superposition processing on the downlink traffic of the original traffic data and the downlink traffic of the alternative traffic data, to obtain the downlink traffic after traffic superposition.
  • the acquiring the network path data of the candidate network path, and performing path evaluation processing on the superimposed traffic data and the network path data to obtain path evaluation data includes: acquiring the candidate path data network path data in the network path data, and perform path evaluation processing on the superimposed traffic data and the network path data to obtain network evaluation data; obtain the current time stamp and the historical prediction data of the candidate network path, and use the current time Delay prediction processing is performed on the stamp and the historical prediction data to obtain target prediction data, so as to determine the network evaluation data and the target prediction data as path evaluation data.
  • the performing delay prediction processing using the current timestamp and the historical prediction data to obtain the target prediction data includes: inputting the current timestamp and the historical prediction data into a pre-trained time delay prediction model, so that the delay prediction model outputs target prediction data.
  • the performing path recommendation processing on the path evaluation data and the network preference data to determine the optimal candidate path of the original network path among the candidate network paths includes: Perform data cleaning processing on the candidate path data to obtain candidate cleaning data, and perform data feature calculation on the candidate cleaning data to obtain candidate data features; obtain feature threshold data in the candidate cleaning data, and Comparing the candidate feature data with the feature threshold data to a candidate comparison result; based on the candidate comparison result, performing route recommendation processing on the route evaluation data and the network preference data, so as to Determine the optimal alternative path of the original network path in the path.
  • the candidate data characteristics include at least one of one-way delay, two-way delay, packet loss rate, and bandwidth utilization.
  • the performing path recommendation processing on the path evaluation data and the network preference data to determine the optimal candidate path of the original network path among the candidate network paths includes: performing data standardization on the path evaluation data to obtain standard evaluation data, and performing path recommendation processing on the standard evaluation data and the network preference data to obtain path recommendation parameters corresponding to the candidate network paths; The degree of recommendation is identified, so as to determine the optimal candidate path of the original network path among the candidate network paths.
  • the recommendation degree is positively correlated with the value of the path recommendation parameter.
  • a processing device for a network path including: a traffic superimposition module configured to acquire original traffic data of the original network path and candidate traffic data of the candidate network path, and performing traffic superposition processing on the original traffic data and the candidate traffic data to obtain the superimposed traffic data; the path evaluation module is configured to obtain the network path data of the candidate network path, and compare the superimposed traffic data and the network performing path evaluation processing on the path data to obtain path evaluation data; the path recommendation module is configured to obtain network preference data corresponding to the path evaluation data, and perform path recommendation processing on the path evaluation data and the network preference data, so as to The best candidate path of the original network path is determined among the candidate network paths.
  • an electronic device including: a processor and a memory; wherein, the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the above-mentioned A method for processing network paths in any exemplary embodiment.
  • a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the network in any of the above-mentioned exemplary embodiments is implemented.
  • the processing method for the path is provided.
  • FIG. 1 schematically shows a schematic flowchart of a method for processing a network path according to an embodiment of the present disclosure
  • FIG. 2 schematically shows a schematic flowchart of a method for acquiring original traffic data and alternative traffic data according to an embodiment of the present disclosure
  • FIG. 3 schematically shows a schematic flowchart of a method for traffic superposition processing according to an embodiment of the present disclosure
  • FIG. 4 schematically shows a schematic flowchart of a method for path evaluation processing according to an embodiment of the present disclosure
  • FIG. 5 schematically shows a schematic flow chart of a method for path recommendation processing according to an embodiment of the present disclosure
  • FIG. 6 schematically shows a schematic flowchart of a method for processing route recommendation according to another embodiment of the present disclosure
  • FIG. 7 schematically shows a system architecture diagram of a network path processing method in an application scenario according to an embodiment of the present disclosure
  • FIG. 8 schematically shows a schematic flowchart of a processing method of a network traffic simulation module according to an embodiment of the present disclosure
  • FIG. 9 schematically shows a schematic flowchart of a method for predicting target prediction data by an alternative path evaluation module according to an embodiment of the present disclosure
  • FIG. 10 schematically shows a schematic flowchart of a processing method of an alternative path recommendation module according to an embodiment of the present disclosure
  • FIG. 11 schematically shows a schematic structural diagram of a network path processing device according to an embodiment of the present disclosure
  • FIG. 12 schematically shows an electronic device for implementing a network path processing method according to an embodiment of the present disclosure
  • Fig. 13 schematically shows a computer-readable storage medium for implementing a network path processing method according to an embodiment of the present disclosure.
  • Example embodiments will now be described more fully with reference to the accompanying drawings.
  • Example embodiments may, however, be embodied in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete and will fully convey the concept of example embodiments to those skilled in the art.
  • the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
  • numerous specific details are provided in order to give a thorough understanding of embodiments of the present disclosure.
  • those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced without one or more of the specific details being omitted, or other methods, components, devices, steps, etc. may be adopted.
  • well-known technical solutions have not been shown or described in detail to avoid obscuring aspects of the present disclosure.
  • the present disclosure provides a network path processing solution, which can more accurately select the optimal alternative path, and comprehensively consider multiple performance indicators that affect path selection, improve network operation and maintenance efficiency from multiple perspectives, and meet the requirements of The path selection needs of users in different situations are more adaptable and flexible.
  • Fig. 1 shows a flow chart of a method for processing a network path according to an embodiment of the present disclosure.
  • the following network path processing method is executed by a network path processing device.
  • the method for processing a network path at least includes the following steps:
  • Step S110 Obtain the original traffic data of the original network path and the candidate traffic data of the candidate network path, and perform traffic superposition processing on the original traffic data and the candidate traffic data to obtain superimposed traffic data.
  • Step S120 Obtain network path data of the candidate network path, and perform path evaluation processing on the superimposed traffic data and network path data to obtain path evaluation data.
  • Step S130 Acquire network preference data corresponding to the path evaluation data, and perform path recommendation processing on the path evaluation data and network preference data, so as to determine the optimal candidate path of the original network path among the candidate network paths.
  • the traffic superposition processing is performed on the original traffic data and the alternative traffic data, and the difference between the original traffic data and the alternative traffic data in the traffic superposition process can be considered, and more accurate
  • the selection of the optimal alternative path improves the accuracy of the determined optimal alternative path; on the other hand, the network preference data and path evaluation data are used for path recommendation processing, and multiple performance indicators that affect path selection are comprehensively considered. Improve the efficiency of network operation and maintenance from multiple angles, and at the same time meet the path selection needs of users in different situations, with stronger adaptability and better flexibility.
  • step S110 the original traffic data of the original network path and the candidate traffic data of the candidate network path are acquired, and traffic superposition processing is performed on the original traffic data and the candidate traffic data to obtain superimposed traffic data.
  • the original network path is a round-trip path used by a service flow during operation in a three-layer network.
  • the alternative network path is a line that can be reselected when the original network path has poor circuit quality or is interrupted.
  • FIG. 2 shows a schematic flowchart of a method for acquiring original traffic data and alternative traffic data. As shown in FIG. 2 , the method at least includes the following steps:
  • step S210 the original route data of the original network route and the candidate route data of the candidate network route are obtained, and data cleaning is performed on the original route data to obtain the original cleaned data.
  • the original path data may include the number of packets sent, the number of packets received, uplink traffic, downlink traffic, bandwidth, packet types, and SLA performance evaluation data of each network router port that the original network path passes through.
  • the SLA performance evaluation data includes maximum packet loss rate, maximum one-way delay, and maximum bandwidth utilization.
  • the original route data may also include other data according to actual conditions and requirements, which is not specifically limited in this exemplary embodiment.
  • the candidate path data may include the number of packets sent, the number of packets received, uplink traffic, downlink traffic, bandwidth, packet types, and SLA performance evaluation data of each network router port that the alternative network path passes through.
  • SLA performance evaluation data includes maximum packet loss rate, maximum one-way delay, and maximum bandwidth utilization.
  • the alternative path data may also include other data according to actual conditions and requirements, which is not specifically limited in this exemplary embodiment.
  • data cleaning may be performed on the obtained original path data to obtain corresponding original cleaning data.
  • step S220 data feature calculation is performed on the original cleaning data to obtain the original data feature, and the original data feature is compared with the original path data to obtain a data comparison result.
  • hourly traffic peak calculations can also be performed to unify the number of received packets and the number of sent packets at an hourly granularity.
  • data feature calculation may also be performed on the original cleaned data to obtain original data features.
  • raw data features and raw route data may be compared.
  • the two-way delay may also be compared with the maximum two-way delay.
  • the data comparison result can be obtained according to the comparison manner.
  • step S230 if the result of the data comparison is that the characteristics of the original data do not meet the requirements of the service level agreement (SLA) performance evaluation data included in the original path data, it is determined that the original network path is abnormal, and the original traffic data and alternatives in the original path data are obtained. Alternative traffic data in path data.
  • SLA service level agreement
  • the interruption of the original network path may also be determined by collecting the interruption information complained by the user, which is not specifically limited in this exemplary embodiment.
  • the original path data may not be used as the threshold value for feature comparison of the original data, and the threshold value may also be set according to historical conditions, for example, the bandwidth utilization threshold value is 80%, and the delay is 10ms etc., which is not specifically limited in this exemplary embodiment.
  • the original flow data in the original path data and the candidate flow data in the candidate path data may be acquired.
  • the original traffic data includes uplink traffic and downlink traffic in the original path data
  • the candidate traffic data includes uplink traffic and downlink traffic in the candidate path data.
  • the abnormality of the original network path is determined through the original path data and the alternative path data, so as to obtain the original traffic data and the alternative traffic data to provide a pre-logic judgment for determining the optimal alternative path, And it provides the theoretical basis and data basis for the subsequent processing.
  • traffic superposition processing may be performed on the original traffic data and the candidate traffic data.
  • FIG. 3 shows a schematic flowchart of a method for traffic superposition processing. As shown in FIG. 3 , the method at least includes the following steps:
  • step S310 the packet type comparison is performed on the original type data in the original route data and the candidate type data in the candidate route data to obtain a type comparison result.
  • the packet type in the original path data is obtained as the original type data
  • the packet type in the alternate path data is obtained as the alternate type data.
  • step S320 if the result of type comparison is that the original type data is the same as the alternative type data, traffic superimposition processing is performed on the original traffic data and the alternative traffic data to obtain superimposed traffic data.
  • the original type data When the original type data is the same as the alternate type data, it indicates that the upstream traffic and downstream traffic of the original path data and the upstream traffic and downstream traffic of the alternate path data are the same, therefore, the original traffic data and the alternate traffic data can be compared Perform traffic superposition processing to obtain superimposed traffic data.
  • the uplink traffic in the original path data and the uplink traffic in the alternative path data may be superimposed, and the downlink traffic in the original path data and the downlink traffic in the alternative path data may be superimposed. Therefore, the superimposed traffic data may include superimposed uplink traffic and superimposed downlink traffic.
  • step S330 if the result of the type comparison is that the original type data is different from the alternative type data, the original route data and the alternative route data are superimposed and updated to obtain superimposed traffic data.
  • the original type data When the original type data is different from the alternative type data, it indicates that the uplink traffic and downlink traffic of the original path data and the uplink traffic and downlink traffic of the alternative path data are different, therefore, the original traffic data and the alternative traffic data can be compared Perform flow update calculations to obtain superimposed flow data.
  • the new service flow the number of new service packets ⁇ the packet length of the alternative path
  • the packet length of the alternative path the traffic volume of the alternative path/the number of packets of the alternative path.
  • the new service traffic is superimposed superimposed traffic data.
  • the alternative path traffic used when calculating the length of the alternative path message can be the upstream or downstream traffic of the alternative network path, and the number of alternative path packets used can be the number of packets sent by the alternative network path or The number of received packets is not specifically limited in this exemplary embodiment.
  • step S120 network path data of the candidate network path is obtained, and path evaluation processing is performed on the superimposed traffic data and network path data to obtain path evaluation data.
  • path evaluation processing may be further performed on the superimposed traffic data.
  • FIG. 4 shows a schematic flowchart of a method for path evaluation processing. As shown in FIG. 4, the method at least includes the following steps:
  • step S410 network path data in the candidate path data is obtained, and path evaluation processing is performed on the superimposed traffic data and network path data to obtain network evaluation data.
  • the network path data in the candidate path data that can be obtained may include the number of routes in the candidate network path, the bandwidth, the total length of link fibers, and other data that can be processed for path evaluation , which is not specifically limited in this exemplary embodiment.
  • path evaluation processing may be performed on overlay traffic data and network path data.
  • Routing hops number of routes in the path - 1
  • Bandwidth utilization total traffic of alternative paths / bandwidth of alternative paths (1)
  • Alternative path length the total length of the link fiber of the alternate path.
  • the routing hops, bandwidth utilization and alternative path lengths are the network evaluation data.
  • step S420 the current time stamp and the historical forecast data of the alternative network path are obtained, and the delay prediction process is performed using the current time stamp and the historical forecast data to obtain the target forecast data, so as to determine the network evaluation data and the target forecast data as path evaluation data.
  • the historical forecast data may include historical traffic data and packet loss rate.
  • the historical traffic data may take the current time as a node and summarize all the traffic data before the current time.
  • the latency prediction process can be performed using the current timestamp and historical prediction data.
  • the current time stamp and historical prediction data are input into a pre-trained delay prediction model, so that the delay prediction model outputs target prediction data.
  • the target prediction data includes one-way delay data and predicted packet loss rate.
  • the one-way delay data is the time difference between sending and receiving service flows.
  • the time difference between sending and receiving service flows the time when the destination end receives the service message - the time when the source end sends the service message.
  • the delay prediction model may be a LightGBM model.
  • the LightGBM algorithm is a machine learning algorithm that implements the idea of GBDT (Gradient Boosting Decision Tree, gradient boosting decision tree).
  • the prediction algorithm formula is:
  • x is the predicted sample
  • T(x; ⁇ m ) represents the decision tree
  • ⁇ m represents the parameters of the decision tree
  • m is the number of trees
  • f m (x) is the predicted value of the sample.
  • the loss function is expressed as:
  • y i is the true value of the i-th sample
  • f m ( xi ) is the predicted value of the i-th sample.
  • the LightGBM model can be trained.
  • the LightGBM model is trained based on the training set, and the parameters of the LightGBM model are adjusted based on the test set.
  • the LightGBM model After the LightGBM model is trained, the current timestamp and historical forecast data can be input into the trained LightGBM model to predict the target forecast data. Therefore, the LightGBM model can output the predicted one-way delay data and packet loss rate as the target forecast data.
  • the target prediction data and network evaluation data can be determined as path evaluation data.
  • path evaluation data can be obtained by performing path evaluation processing on superimposed traffic data and network path data, which provides a data basis for determining the optimal alternative path, so that the determination of the optimal alternative path can synthesize multiple performance, which ensures the determination of the optimal alternative path and the efficiency of network operation and maintenance.
  • step S130 network preference data corresponding to the path evaluation data is obtained, and path recommendation processing is performed on the path evaluation data and network preference data, so as to determine the optimal candidate path of the original network path among the candidate network paths.
  • user preferences may be considered comprehensively, so network preference data may be further acquired.
  • path recommendation processing can be performed on the path evaluation data and the network preference data.
  • FIG. 5 shows a schematic flowchart of a method for route recommendation processing. As shown in FIG. 5, the method at least includes the following steps:
  • step S510 data cleaning processing is performed on the candidate path data to obtain candidate cleaning data, and data feature calculation is performed on the candidate cleaning data to obtain candidate data features.
  • the data cleaning process may include unit unification of candidate path data, missing value processing, outlier value processing, etc., and may also include other processing methods, which are not specifically limited in this exemplary embodiment.
  • hourly traffic peak calculations can also be performed to unify the number of received packets and the number of sent packets at an hourly granularity.
  • data feature calculation may also be performed on the candidate cleaning data to obtain candidate data features.
  • the candidate data characteristics may include one-way delay, two-way delay, packet loss rate, and bandwidth utilization.
  • step S520 the characteristic threshold data in the candidate cleaning data is obtained, and the candidate characteristic data and the characteristic threshold data are compared to a candidate comparison result.
  • the feature threshold data obtained from the candidate cleaning data may include the SLA performance evaluation data of the candidate network path, that is, the maximum packet loss rate, the maximum one-way delay and the maximum bandwidth utilization.
  • the candidate feature data can be compared to feature threshold data.
  • a ratio between the candidate characteristic data and the characteristic threshold data may be calculated to obtain the ratio of the two as the candidate comparison result.
  • step S530 based on the alternative comparison results, path recommendation processing is performed on the path evaluation data and network preference data, so as to determine the optimal alternative path of the original network path among the alternative network paths.
  • the path recommendation process may continue to be performed on the path evaluation data and network preference data of the candidate network path.
  • the candidate comparison result is that the ratio of the candidate feature data to the feature threshold data is greater than 1
  • the candidate network path can be discarded, and path recommendation processing is performed on the path evaluation data and network preference data of other candidate network paths.
  • FIG. 6 shows a schematic flowchart of a method for further processing path recommendation. As shown in FIG. 6, the method at least includes the following steps:
  • step S610 data standardization is performed on the path evaluation data to obtain standard evaluation data, and path recommendation processing is performed on the standard evaluation data and network preference data to obtain path recommendation parameters corresponding to candidate network paths.
  • the Z-score algorithm can be used to standardize the path evaluation data:
  • ⁇ is the path evaluation data of multiple candidate network paths, that is, the corresponding average value of routing hops, bandwidth utilization, candidate path length and one-way delay data.
  • ⁇ is the corresponding standard deviation of the path evaluation data of multiple candidate network paths.
  • ⁇ is the mean value of the candidate path lengths of the 5 candidate network paths
  • ⁇ is the standard deviation of the candidate path lengths of the 5 candidate network paths
  • the network preference data of the alternative path length, routing hops, one-way delay data, bandwidth utilization rate and packet loss rate in the standard evaluation data are w 1 , w 2 , w 3 , w 4 and w 5 respectively. Moreover, since the smaller the five indexes are, the path is optimal, so the five network preference data are all less than 0.
  • the network preference data when performing path recommendation processing on the standard evaluation data and network preference data, can be used as the weight for calculation:
  • w 1 +w 2 +w 3 +w 4 +w 5 -1, and w 1 ⁇ 0, w 2 ⁇ 0, w 3 ⁇ 0, w 4 ⁇ 0, w 5 ⁇ 0, ⁇ 1 is more The mean value of the candidate path lengths of the candidate network paths, ⁇ 1 is the standard deviation of the candidate path lengths of the multiple candidate network paths, ⁇ 2 is the mean value of the routing hops of the multiple candidate network paths, and ⁇ 2 is The standard deviation of the routing hops of multiple alternative network paths, ⁇ 3 is the mean value of the one-way delay data of multiple alternative network paths, and ⁇ 3 is the standard deviation of the one-way delay data of multiple alternative network paths , ⁇ 4 is the average value of the bandwidth utilization ratio of multiple alternative network paths, ⁇ 4 is the standard deviation of the bandwidth utilization ratio of multiple alternative network paths, ⁇ 5 is the average value of the packet loss rate of multiple alternative network paths, ⁇ 5 is the standard deviation of the packet loss rate of multiple alternative network paths.
  • Y is a
  • step S620 the recommendation degree identification is performed on the path recommendation parameters, so as to determine the optimal candidate path of the original network path among the candidate network paths.
  • the path recommendation parameters of multiple candidate network paths can be compared to determine the candidate network path with the largest path recommendation parameter as the optimal candidate path of the original network path.
  • combining the user's network preference data and path evaluation data to perform path recommendation processing to determine the optimal alternative path can not only meet user needs, but also further improve network operation and maintenance efficiency, and the recommendation effect is better.
  • Figure 7 shows a system architecture diagram of a network path processing method in an application scenario.
  • the system architecture includes a client device, a router, a customer service center, a data collection module, a data storage module, a data processing module, An end-to-end service quality assessment module, a network flow simulation module, an alternative path evaluation module and an alternative path recommendation module.
  • the client device is a client-side edge network device.
  • the router is an IP (Internet Protocol, Internet Interconnection Protocol) network router device.
  • the customer business center is a cloud server cluster that provides services for customers or operators.
  • the data collection module can collect the original path data of the original network path and the candidate path data of the candidate network path.
  • the original path data may include the number of packets sent, the number of packets received, upstream traffic, downstream traffic, bandwidth, packet types, and SLA performance evaluation data of each network router port that the original network path passes through.
  • the SLA performance evaluation data Including maximum packet loss rate, maximum one-way delay and maximum bandwidth utilization.
  • the original route data may also include other data according to actual conditions and requirements, which is not specifically limited in this exemplary embodiment.
  • the candidate path data may include the number of packets sent, the number of packets received, uplink traffic, downlink traffic, bandwidth, packet types, and SLA performance evaluation data of each network router port that the alternative network path passes through.
  • SLA performance evaluation data includes maximum packet loss rate, maximum one-way delay, and maximum bandwidth utilization.
  • the candidate path data may also include other data according to actual conditions and requirements, which is not specifically limited in this exemplary embodiment.
  • the data storage module can efficiently and quickly store a large amount of collected original path data and alternative path data.
  • data cleaning may be performed on the obtained original path data and candidate path data to obtain corresponding original cleaning data and candidate cleaning data.
  • the data cleaning process may include unit unification of various performance data, missing value processing, outlier value processing, etc., and may also include other processing methods, which are not specifically limited in this exemplary embodiment.
  • the data processing module can process the collected original cleaning data and alternative cleaning data.
  • hourly traffic peak calculation can also be performed to unify the number of received packets and sent packets according to hourly granularity.
  • data feature calculation may be performed on the original cleaned data to obtain original data features, and data feature calculation may be performed on candidate cleaned data to obtain candidate data features.
  • both the original data feature and the candidate data feature may include one-way delay, two-way delay, packet loss rate, and bandwidth utilization rate.
  • the end-to-end service quality evaluation module uses the end-to-end service quality evaluation method to evaluate the end-to-end service quality.
  • the characteristics of the original data can be compared with the original path data to evaluate three data comparison results: normal, poor quality and interruption.
  • the two-way delay may also be compared with the maximum two-way delay.
  • the data comparison result can be obtained according to the comparison manner.
  • the packet loss rate is 100%, it can be determined that the original network path is interrupted, so the original network path is abnormal.
  • the packet loss rate is not 100%, but other original data characteristics exceed the original path data, it is determined that the original network path has poor quality, and therefore the original network path is abnormal.
  • the interruption of the original network path may also be determined by collecting interruption information complained by users, which is not specifically limited in this exemplary embodiment.
  • the original path data may not be used as the threshold value for feature comparison of the original data, and the threshold value may also be set according to historical conditions, for example, the bandwidth utilization threshold value is 80%, and the delay is 10ms etc., which is not specifically limited in this exemplary embodiment.
  • the original flow data in the original path data and the candidate flow data in the candidate path data may be acquired.
  • the original traffic data includes uplink traffic and downlink traffic in the original path data
  • the candidate traffic data includes uplink traffic and downlink traffic in the candidate path data.
  • the network traffic simulation module Before evaluating and recommending alternative paths, the network traffic simulation module needs to judge whether the packet type currently carried by the alternative network path is the same as that of the original network path, and then superimpose the original network path on the basis of the alternative network path The flow data of the alternative path network flow simulation is carried out.
  • Figure 8 shows a schematic flow chart of the processing method of the network traffic simulation module, as shown in Figure 8:
  • step S810 it is determined whether the message types are the same.
  • the packet type comparison is performed on the original type data in the original route data and the alternative type data in the alternative route data to obtain a type comparison result.
  • the packet type in the original path data is obtained as the original type data
  • the packet type in the alternate path data is obtained as the alternate type data. Further, it can be judged whether the original type data is the same as the alternative type data, so as to obtain a type comparison result.
  • the original traffic data and the alternative traffic data are subjected to traffic superposition processing to obtain superimposed traffic data.
  • the original type data When the original type data is the same as the alternate type data, it indicates that the upstream traffic and downstream traffic of the original path data and the upstream traffic and downstream traffic of the alternate path data are the same, therefore, the original traffic data and the alternate traffic data can be compared Perform traffic superposition processing to obtain superimposed traffic data.
  • the uplink traffic in the original path data and the uplink traffic in the alternative path data may be superimposed, and the downlink traffic in the original path data and the downlink traffic in the alternative path data may be superimposed. Therefore, the superimposed traffic data may include superimposed uplink traffic and superimposed downlink traffic.
  • step S830 the length of the candidate line message is calculated.
  • the original path data and the alternate path data are superimposed and updated to obtain superimposed traffic data.
  • the original type data When the original type data is different from the alternative type data, it indicates that the uplink traffic and downlink traffic of the original path data and the uplink traffic and downlink traffic of the alternative path data are different, therefore, the original traffic data and the alternative traffic data can be compared Perform flow update calculations to obtain superimposed flow data.
  • the packet length of the alternative path may be calculated first.
  • the packet length of the alternative path traffic of the alternate path/number of packets of the alternate path.
  • step S840 the traffic of the service on the alternative line is calculated.
  • the new service flow After calculating the packet length of the alternative path, the new service flow can be calculated.
  • the new service flow number of new service packets ⁇ length of alternative path packets.
  • the new service traffic is superimposed superimposed traffic data.
  • the alternative path traffic used when calculating the length of the alternative path message can be the upstream or downstream traffic of the alternative network path, and the number of alternative path packets used can be the number of packets sent by the alternative network path or The number of received packets is not specifically limited in this exemplary embodiment.
  • traffic overlay processing is done when the original type data is different from the alternative type data.
  • the alternative path evaluation module calculates the bandwidth utilization rate of the alternative network path and evaluates the time difference between sending and receiving service flows after the alternative network path updates the traffic.
  • path evaluation processing may be further performed on the superimposed traffic data.
  • the network path data in the candidate path data that can be obtained may include the number of routes in the candidate network path, the bandwidth, the total length of link fibers, and other data that can be processed for path evaluation , which is not specifically limited in this exemplary embodiment.
  • path evaluation processing may be performed on overlay traffic data and network path data.
  • the number of route hops the number of routes in the path-1
  • the bandwidth utilization rate the total traffic of the alternative path/the bandwidth of the alternative path
  • the length of the alternative path the total length of link fibers of the alternative path.
  • the route hop count, bandwidth utilization rate and candidate path length are the network evaluation data.
  • Fig. 9 shows a schematic flowchart of a method for predicting target prediction data by the alternative path evaluation module, as shown in Fig. 9:
  • step S910 the historical traffic, time stamp, one-way delay and packet loss rate of the alternative path are input.
  • the historical forecast data may include historical traffic data and packet loss rate.
  • the historical traffic data may take the current time as a node and summarize all the traffic data before the current time.
  • the latency prediction process can be performed using the current timestamp and historical prediction data.
  • the current time stamp and historical prediction data are input into a pre-trained delay prediction model, so that the delay prediction model outputs target prediction data.
  • the target prediction data includes one-way delay data and predicted packet loss rate.
  • the one-way delay data is the time difference between sending and receiving service flows.
  • the time difference between sending and receiving service flows the time when the destination end receives the service message - the time when the source end sends the service message.
  • the delay prediction model may be a LightGBM model. For example, obtain historical traffic samples, time stamp samples, one-way delay samples, and packet loss rate samples of alternative network paths.
  • step S920 it is divided into training set or test set.
  • the sample data is divided into training set and test set according to the ratio of 7:3.
  • step S930 the LightGBM algorithm model is trained based on the training set.
  • the LightGBM model is trained based on the training set, and the parameters of the LightGBM model are adjusted based on the test set.
  • step S940 the one-way time delay and packet loss rate of the alternative path after superimposing traffic are predicted.
  • the LightGBM model After the LightGBM model is trained, the current timestamp and historical forecast data can be input into the trained LightGBM model to predict the target forecast data. Therefore, the LightGBM model can output the predicted one-way delay data and packet loss rate as the target forecast data.
  • the target prediction data and network evaluation data can be determined as path evaluation data.
  • the alternative path recommendation module recommends alternative network paths based on the minimum bandwidth utilization, the minimum routing hops, the shortest time, the shortest path length, the minimum packet loss rate, and the comprehensive indicators of the five, so that users can choose according to their own preferences and Select an alternative network path based on business requirements.
  • Fig. 10 shows a schematic flow chart of the processing method of the alternative path recommendation module, as shown in Fig. 10:
  • step S1010 path length, route hop count, one-way delay, bandwidth utilization rate and packet loss rate are calculated.
  • step S1020 a ratio calculation is performed with the SLA maximum bandwidth utilization rate, the maximum one-way delay and the maximum packet loss rate.
  • the data cleaning process may include unit unification of candidate path data, missing value processing, outlier value processing, etc., and may also include other processing methods, which are not specifically limited in this exemplary embodiment.
  • hourly traffic peak calculations can also be performed to unify the number of received packets and the number of sent packets at an hourly granularity.
  • the candidate data characteristics may include one-way delay, two-way delay, packet loss rate, and bandwidth utilization.
  • the candidate feature data can be compared to feature threshold data.
  • the characteristic threshold data obtained from the candidate cleaning data may include the SLA performance evaluation data of the candidate network path, that is, the maximum packet loss rate, the maximum one-way delay and the maximum bandwidth utilization.
  • the candidate network path can be discarded, and path recommendation processing is performed on the path evaluation data and network preference data of other candidate network paths.
  • step S1040 calculate the Z-score value.
  • the path recommendation process may continue to be performed on the path evaluation data and network preference data of the candidate network path.
  • Data standardization is performed on the path evaluation data to obtain standard evaluation data
  • path recommendation processing is performed on the standard evaluation data and network preference data to obtain path recommendation parameters corresponding to candidate network paths.
  • the Z-score algorithm may be used to standardize the path evaluation data according to formula (5) to obtain standard evaluation data.
  • ⁇ is the mean value of the candidate path lengths of the 5 candidate network paths
  • ⁇ is the standard deviation of the candidate path lengths of the 5 candidate network paths
  • step S1050 statistics are made on the percentages of preferences of the users for the five indicators.
  • network preference data can also be further obtained.
  • the network preference data is the user's preference data on the route hop count, bandwidth utilization rate, candidate path length, one-way delay data and packet loss rate in the path evaluation data.
  • the network preference data of the alternative path length, routing hops, one-way delay data, bandwidth utilization rate and packet loss rate in the standard evaluation data are w 1 , w 2 , w 3 , w 4 and w 5 respectively.
  • the five network preference data are all less than 0.
  • the path recommendation parameter y can be obtained by calculating with the network preference data as a weight according to formula (6).
  • step S1070 the number of the candidate path with the largest value is output.
  • the path recommendation parameters of multiple candidate network paths can be compared to determine the candidate network path with the largest path recommendation parameter as the optimal candidate path of the original network path.
  • an example may also be used to illustrate the processing method of the network path.
  • the packet type of the original network path is IPV6 (Internet Protocol Version 6, Internet Protocol Version 6), there is no packet in the alternative network path 1, and the packet type in the alternate network path 2 is IPV6,
  • the packet type in the candidate network path 3 is SRV6.
  • the SLA performance evaluation data Including maximum packet loss rate, maximum one-way delay and maximum bandwidth utilization.
  • the service path preference data of the user that is, network preference data, may also be obtained.
  • hourly traffic peak calculations can also be performed to unify the number of received packets and the number of sent packets at an hourly granularity.
  • End-to-end based on performance data (packet loss rate, one-way delay, two-way delay, and bandwidth utilization) and SLA data (maximum packet loss, maximum one-way delay, maximum two-way delay, and maximum bandwidth utilization) Business quality assessment.
  • the original traffic data and the alternative traffic data are subjected to traffic superposition processing to obtain superimposed traffic data.
  • the original path data and the alternate path data are superimposed and updated to obtain superimposed traffic data.
  • the candidate network path 3 is finally selected as the optimal candidate path of the original network path.
  • the network path processing method in this application scenario performs traffic superposition processing on the original traffic data and the alternative traffic data, and considers the difference between the original traffic data and the alternative traffic data in the traffic superposition process, which can be more accurate
  • the selection of the optimal alternative path improves the accuracy of the determined optimal alternative path; on the other hand, the network preference data and path evaluation data are used for path recommendation processing, and multiple performance indicators that affect path selection are comprehensively considered. , improve the efficiency of network operation and maintenance from multiple perspectives, and at the same time meet the path selection needs of users in different situations, with stronger adaptability and better flexibility.
  • FIG. 11 shows a schematic structural diagram of a network path processing device.
  • the network path processing device 1100 may include: a traffic superposition module 1110 , a path evaluation module 1120 and a path recommendation module 1130 . in:
  • the traffic superposition module 1110 is configured to obtain the original traffic data of the original network path and the candidate traffic data of the candidate network path, and perform traffic superposition processing on the original traffic data and the candidate traffic data to obtain the superimposed traffic data;
  • the path evaluation module 1120 configured to acquire network path data of alternative network paths, and perform path evaluation processing on the superimposed traffic data and network path data to obtain path evaluation data;
  • the path recommendation module 1130 is configured to obtain network preference data corresponding to the path evaluation data, Path recommendation processing is performed on the path evaluation data and network preference data to determine the optimal alternative path of the original network path among the alternative network paths.
  • the management configuration information includes: the acquisition of the original traffic data of the original network path and the alternative traffic data of the alternative network path includes:
  • the result of the data comparison is that the characteristics of the original data do not meet the requirements of the service level agreement (SLA) performance evaluation data included in the original path data, determine that the original network path is abnormal, and obtain the original path data in the original path. Alternate flow data in flow data and alternative path data.
  • SLA service level agreement
  • performing traffic superposition processing on the original traffic data and the candidate traffic data to obtain superimposed traffic data includes:
  • the type comparison result is that the original type data is the same as the alternative type data, performing traffic superposition processing on the original traffic data and the alternative traffic data to obtain superimposed traffic data;
  • the acquiring the network path data of the candidate network path, and performing path evaluation processing on the superimposed traffic data and the network path data to obtain path evaluation data includes:
  • the one-way delay data obtained by using the current time stamp and the historical traffic data to perform delay prediction processing includes:
  • the current time stamp and the historical prediction data are input into a pre-trained delay prediction model, so that the delay prediction model outputs target prediction data.
  • the path recommendation process is performed on the path evaluation data and the network preference data, so as to determine the optimal alternative of the original network path among the candidate network paths.
  • Choose a path including:
  • the path recommendation process is performed on the path evaluation data and the network preference data, so as to determine the optimal alternative of the original network path among the candidate network paths.
  • Choose a path including:
  • the recommendation degree identification is carried out on the path recommendation parameters, so as to determine the optimal candidate path of the original network path among the candidate network paths.
  • modules or units of the network path processing apparatus 1100 are mentioned in the above detailed description, this division is not mandatory. Actually, according to the embodiment of the present disclosure, the features and functions of two or more modules or units described above may be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided to be embodied by a plurality of modules or units.
  • FIG. 12 An electronic device 1200 according to such an embodiment of the present invention is described below with reference to FIG. 12 .
  • the electronic device 1200 shown in FIG. 12 is only an example, and should not limit the functions and scope of use of this embodiment of the present invention.
  • electronic device 1200 takes the form of a general-purpose computing device.
  • the components of the electronic device 1200 may include, but are not limited to: at least one processing unit 1210, at least one storage unit 1220, a bus 1230 connecting different system components (including the storage unit 1220 and the processing unit 1210), and a display unit 1240.
  • the storage unit stores program codes, and the program codes can be executed by the processing unit 1210, so that the processing unit 1210 executes various exemplary methods according to the present invention described in the "Exemplary Methods" section of this specification. Example steps.
  • the storage unit 1220 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 1221 and/or a cache storage unit 1222 , and may further include a read-only storage unit (ROM) 1223 .
  • RAM random access storage unit
  • ROM read-only storage unit
  • Storage unit 1220 may also include a program/utility 1224 having a set (at least one) of program modules 1225, such program modules 1225 including but not limited to: an operating system, one or more application programs, other program modules, and program data, Implementations of networked environments may be included in each or some combination of these examples.
  • program modules 1225 including but not limited to: an operating system, one or more application programs, other program modules, and program data, Implementations of networked environments may be included in each or some combination of these examples.
  • Bus 1230 may represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local area using any of a variety of bus structures. bus.
  • the electronic device 1200 can also communicate with one or more external devices 1400 (such as keyboards, pointing devices, Bluetooth devices, etc.), and can also communicate with one or more devices that enable the user to interact with the electronic device 1200, and/or communicate with Any device (eg, router, modem, etc.) that enables the electronic device 1200 to communicate with one or more other computing devices. Such communication may occur through input/output (I/O) interface 1250 .
  • the electronic device 1200 can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN) and/or a public network such as the Internet) through the network adapter 1260 .
  • networks such as a local area network (LAN), a wide area network (WAN) and/or a public network such as the Internet
  • the network adapter 1240 communicates with other modules of the electronic device 1200 through the bus 1230 .
  • other hardware and/or software modules may be used in conjunction with electronic device 1200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives And data backup storage system, etc.
  • the exemplary embodiments described here can be implemented by software, or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of software products, and the software products can be stored in a non-volatile storage medium (which can be CD-ROM, U disk, mobile hard disk, etc.) or on the network , including several instructions to make a computing device (which may be a personal computer, a server, a terminal device, or a network device, etc.) execute the method according to the embodiment of the present disclosure.
  • a computing device which may be a personal computer, a server, a terminal device, or a network device, etc.
  • a computer-readable storage medium on which a program product capable of implementing the above-mentioned method in this specification is stored.
  • various aspects of the present invention can also be implemented in the form of a program product, which includes program code, and when the program product is run on a terminal device, the program code is used to make the The terminal device executes the steps according to various exemplary embodiments of the present invention described in the "Exemplary Method" section above in this specification.
  • a program product 1300 for realizing the above method according to an embodiment of the present invention is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program codes, and can be used in terminal equipment, For example running on a personal computer.
  • CD-ROM compact disk read-only memory
  • the program product of the present invention is not limited thereto.
  • a readable storage medium may be any tangible medium containing or storing a program, and the program may be used by or in combination with an instruction execution system, apparatus or device.
  • the program product may reside on any combination of one or more readable media.
  • the readable medium may be a readable signal medium or a readable storage medium.
  • the readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: electrical connection with one or more conductors, portable disk, hard disk, random access memory (RAM), read only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
  • a computer readable signal medium may include a data signal carrying readable program code in baseband or as part of a carrier wave. Such propagated data signals may take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.
  • a readable signal medium may also be any readable medium other than a readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.
  • Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
  • the remote computing device may be connected to the user computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., using an Internet service provider). business to connect via the Internet).
  • LAN local area network
  • WAN wide area network
  • Internet service provider an Internet service provider

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The present disclosure relates to the technical field of network communications, and provides a network path processing method and apparatus, a storage medium, and an electronic device. The network path processing method comprises: obtaining original traffic data of an original network path and alternative traffic data of alternative network paths, and performing traffic superposition processing on the original traffic data and the alternative traffic data to obtain superimposed traffic data; obtaining network path data of the alternative network paths, and performing path evaluation processing on the superimposed traffic data and network path data to obtain path evaluation data; obtaining network preference data corresponding to the path evaluation data, and performing path recommendation processing on the path evaluation data and the network preference data to determine an optimal alternative path of the original network path from among the alternative network paths.

Description

网络路径的处理方法及装置、存储介质、电子设备Network path processing method and device, storage medium, and electronic equipment
相关申请的交叉引用Cross References to Related Applications
本申请是以CN申请号为202110924959.4，申请日为2021年8月12日的申请为基础，并主张其优先权，该CN申请的公开内容在此作为整体引入本申请中。This application is based on the application with CN application number 202110924959.4 and the application date is August 12, 2021, and claims its priority. The disclosure content of this CN application is hereby incorporated into this application as a whole.
技术领域technical field
本公开涉及网络通信技术领域，尤其涉及一种网络路径的处理方法及装置、存储介质、电子设备。The present disclosure relates to the technical field of network communication, and in particular, to a network path processing method and device, a storage medium, and electronic equipment.
背景技术Background technique
在三层网络中，业务流在运行过程中通常会进行往返路径的选取。如果出现电路质差或者中断，会重新选择线路。基于5G(5th Generation Mobile Communication Technology，第五代移动通信技术)移动承载业务的SLA(Service Level Agreement，服务级别协议)保障和运维诉求，结合现有性能检测技术的优缺点，创新的iFIT(in-situ Flow Information Telemetry，随流检测)方案应运而生。该方案将基于端到端的随流检测模式分析业务质差或者中断的情况，进而转换成逐跳的检测模式，对电路问题点进行定位，并重新规划路线。In a three-layer network, the service flow usually selects a round-trip path during operation. If the circuit quality is poor or interrupted, the line will be reselected. Based on the SLA (Service Level Agreement, Service Level Agreement) guarantee and operation and maintenance requirements of 5G (5th Generation Mobile Communication Technology) mobile bearer services, combined with the advantages and disadvantages of existing performance detection technologies, the innovative iFIT ( The in-situ Flow Information Telemetry (Flow Information Telemetry) scheme came into being. This solution will analyze the poor or interrupted service quality based on the end-to-end follow-the-flow detection mode, and then switch to the hop-by-hop detection mode to locate circuit problems and re-plan routes.
针对重新规划的路线，需要在原始业务性能数据基础上增加新业务的性能数据，并基于客户喜好以及网络性能进行备选路径预判，以实现业务流的高效运转，进而提升用户感知。目前有段路由策略的路径选择方法通过叠加备选链路代价选择最优备选路径，也有基于负载均衡与QoE(Quality of Experience，体验质量)度量值的SDN(Software Defined Network，软件定义网络)网络路径选择方法，为每条已确定的路径计算其QoE的度量值，最终选择具有最大QoE度量值的路径为输出路径，还有一种网络路径选择方法是基于最小链路长度进行路径的选取。For the replanned route, it is necessary to add new service performance data on the basis of the original service performance data, and predict alternative paths based on customer preferences and network performance, so as to achieve efficient operation of service flows and improve user perception. At present, there are path selection methods based on segment routing strategies to select the optimal alternative path by superimposing the cost of alternative links, and there are also SDN (Software Defined Network) based on load balancing and QoE (Quality of Experience) metrics. The network path selection method calculates its QoE metric value for each determined path, and finally selects the path with the largest QoE metric value as the output path. Another network path selection method is to select a path based on the minimum link length.
发明内容Contents of the invention
根据本公开实施例的第一个方面，提供一种网络路径的处理方法，包括：获取原始网络路径的原始流量数据和备选网络路径的备选流量数据，对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据；获取所述备选网络路径的网 络路径数据，并对所述叠加流量数据和所述网络路径数据进行路径评估处理得到路径评估数据；获取与所述路径评估数据对应的网络偏好数据，对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径。According to the first aspect of the embodiments of the present disclosure, there is provided a method for processing a network path, including: acquiring original traffic data of the original network path and candidate traffic data of the candidate network path, and processing the original traffic data and the The alternative traffic data is subjected to traffic superposition processing to obtain superimposed traffic data; the network path data of the candidate network path is obtained, and path evaluation processing is performed on the superimposed traffic data and the network path data to obtain path evaluation data; network preference data corresponding to the path evaluation data, and perform path recommendation processing on the path evaluation data and the network preference data, so as to determine the optimal candidate path of the original network path among the candidate network paths.
在一些实施例中，所述获取原始网络路径的原始流量数据和备选网络路径的备选流量数据包括：获取原始网络路径的原始路径数据和备选网络路径的备选路径数据，并对所述原始路径数据进行数据清洗处理得到原始清洗数据；对所述原始清洗数据进行数据特征计算得到原始数据特征，并对所述原始数据特征和所述原始路径数据进行比较得到数据比较结果；若所述数据比较结果为所述原始数据特征未满足所述原始路径数据中包括的服务级别协议SLA性能评估数据的要求，确定所述原始网络路径异常，并获取所述原始路径数据中的原始流量数据和备选路径数据中的备选流量数据。In some embodiments, the obtaining the original flow data of the original network path and the candidate flow data of the candidate network path includes: acquiring the original path data of the original network path and the candidate path data of the candidate network path, and Perform data cleaning processing on the original path data to obtain original cleaning data; perform data feature calculation on the original cleaning data to obtain original data features, and compare the original data features with the original path data to obtain a data comparison result; if the The result of the data comparison is that the characteristics of the original data do not meet the requirements of the service level agreement SLA performance evaluation data included in the original path data, determine that the original network path is abnormal, and obtain the original traffic data in the original path data and alternative flow data in the alternative path data.
在一些实施例中，所述原始数据特征包括单向时延、双向时延、丢包率和带宽利用率中的至少一项。In some embodiments, the raw data characteristics include at least one of one-way delay, two-way delay, packet loss rate, and bandwidth utilization rate.
在一些实施例中，在所述原始数据特征包括所述单向时延的情况下，所述SLA性能评估数据包括最大单向时延；在所述原始数据特征包括所述双向时延的情况下，所述SLA性能评估数据包括最大双向时延；在所述原始数据特征包括所述丢包率的情况下，所述SLA性能评估数据包括最大丢包率；在所述原始数据特征包括所述带宽利用率的情况下，所述SLA性能评估数据包括最大带宽利用率。In some embodiments, when the original data feature includes the one-way delay, the SLA performance evaluation data includes the maximum one-way delay; when the original data feature includes the two-way delay In this case, the SLA performance evaluation data includes the maximum two-way delay; when the original data feature includes the packet loss rate, the SLA performance evaluation data includes the maximum packet loss rate; when the original data feature includes the packet loss rate In the case of the above-mentioned bandwidth utilization ratio, the SLA performance evaluation data includes the maximum bandwidth utilization ratio.
在一些实施例中，所述对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据包括：对所述原始路径数据中的原始类型数据和所述备选路径数据中的备选类型数据进行报文类型比较得到类型比较结果；若所述类型比较结果为所述原始类型数据与所述备选类型数据相同，对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据；若所述类型比较结果为所述原始类型数据与所述备选类型数据不同，对所述原始路径数据和所述备选路径数据进行叠加更新计算得到叠加流量数据。In some embodiments, performing traffic superposition processing on the original traffic data and the candidate traffic data to obtain the superimposed traffic data includes: The alternative type data is compared to the message type to obtain the type comparison result; if the type comparison result is that the original type data is the same as the alternative type data, the original traffic data and the alternative traffic data are performed Flow superposition processing to obtain superimposed traffic data; if the type comparison result is that the original type data is different from the alternative type data, superimposing and updating the original path data and the alternative path data to obtain superimposed traffic data .
在一些实施例中，对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据包括：对所述原始流量数据的上行流量和所述备选流量数据的上行流量进行流量叠加处理，得到流量叠加后的上行流量；对所述原始流量数据的下行流量和所述备选流量数据的下行流量进行流量叠加处理，得到流量叠加后的下行流量。In some embodiments, performing traffic superposition processing on the original traffic data and the alternative traffic data to obtain superimposed traffic data includes: performing traffic on the upstream traffic of the original traffic data and the upstream traffic of the alternative traffic data superposition processing, to obtain the upstream traffic after traffic superimposition; perform traffic superposition processing on the downlink traffic of the original traffic data and the downlink traffic of the alternative traffic data, to obtain the downlink traffic after traffic superposition.
在一些实施例中，所述获取所述备选网络路径的网络路径数据，并对所述叠加流 量数据和所述网络路径数据进行路径评估处理得到路径评估数据包括：获取所述备选路径数据中的网络路径数据，并对所述叠加流量数据和所述网络路径数据进行路径评估处理得到网络评估数据；获取当前时间戳和所述备选网络路径的历史预测数据，并利用所述当前时间戳和所述历史预测数据进行时延预测处理得到目标预测数据，以确定所述网络评估数据和所述目标预测数据为路径评估数据。In some embodiments, the acquiring the network path data of the candidate network path, and performing path evaluation processing on the superimposed traffic data and the network path data to obtain path evaluation data includes: acquiring the candidate path data network path data in the network path data, and perform path evaluation processing on the superimposed traffic data and the network path data to obtain network evaluation data; obtain the current time stamp and the historical prediction data of the candidate network path, and use the current time Delay prediction processing is performed on the stamp and the historical prediction data to obtain target prediction data, so as to determine the network evaluation data and the target prediction data as path evaluation data.
在一些实施例中，所述利用所述当前时间戳和所述历史预测数据进行时延预测处理得到目标预测数据包括：将所述当前时间戳和所述历史预测数据输入至预先训练好的时延预测模型中，以使所述时延预测模型输出目标预测数据。In some embodiments, the performing delay prediction processing using the current timestamp and the historical prediction data to obtain the target prediction data includes: inputting the current timestamp and the historical prediction data into a pre-trained time delay prediction model, so that the delay prediction model outputs target prediction data.
在一些实施例中，所述对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径包括：对所述备选路径数据进行数据清洗处理得到备选清洗数据，并对所述备选清洗数据进行数据特征计算得到备选数据特征；获取所述备选清洗数据中的特征门限数据，并对所述备选特征数据和所述特征门限数据进行比较到备选比较结果；基于所述备选比较结果，对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径。In some embodiments, the performing path recommendation processing on the path evaluation data and the network preference data to determine the optimal candidate path of the original network path among the candidate network paths includes: Perform data cleaning processing on the candidate path data to obtain candidate cleaning data, and perform data feature calculation on the candidate cleaning data to obtain candidate data features; obtain feature threshold data in the candidate cleaning data, and Comparing the candidate feature data with the feature threshold data to a candidate comparison result; based on the candidate comparison result, performing route recommendation processing on the route evaluation data and the network preference data, so as to Determine the optimal alternative path of the original network path in the path.
在一些实施例中，所述备选数据特征包括单向时延、双向时延、丢包率和带宽利用率中的至少一项。In some embodiments, the candidate data characteristics include at least one of one-way delay, two-way delay, packet loss rate, and bandwidth utilization.
在一些实施例中，所述对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径包括：对所述路径评估数据进行数据标准化得到标准评估数据，并对所述标准评估数据和所述网络偏好数据进行路径推荐处理得到与所述备选网络路径对应的路径推荐参数；对所述路径推荐参数进行推荐程度识别，以在所述备选网络路径中确定所述原始网络路径的最优备选路径。In some embodiments, the performing path recommendation processing on the path evaluation data and the network preference data to determine the optimal candidate path of the original network path among the candidate network paths includes: performing data standardization on the path evaluation data to obtain standard evaluation data, and performing path recommendation processing on the standard evaluation data and the network preference data to obtain path recommendation parameters corresponding to the candidate network paths; The degree of recommendation is identified, so as to determine the optimal candidate path of the original network path among the candidate network paths.
在一些实施例中，所述推荐程度与所述路径推荐参数的数值呈正相关关系。In some embodiments, the recommendation degree is positively correlated with the value of the path recommendation parameter.
根据本公开实施例的第二个方面，提供一种网络路径的处理装置，包括：流量叠加模块，被配置为获取原始网络路径的原始流量数据和备选网络路径的备选流量数据，对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据；路径评估模块，被配置为获取所述备选网络路径的网络路径数据，并对所述叠加流量数据和所述网络路径数据进行路径评估处理得到路径评估数据；路径推荐模块，被配置为获取与所述路径评估数据对应的网络偏好数据，对所述路径评估数据和所述网络偏好 数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径。According to a second aspect of an embodiment of the present disclosure, there is provided a processing device for a network path, including: a traffic superimposition module configured to acquire original traffic data of the original network path and candidate traffic data of the candidate network path, and performing traffic superposition processing on the original traffic data and the candidate traffic data to obtain the superimposed traffic data; the path evaluation module is configured to obtain the network path data of the candidate network path, and compare the superimposed traffic data and the network performing path evaluation processing on the path data to obtain path evaluation data; the path recommendation module is configured to obtain network preference data corresponding to the path evaluation data, and perform path recommendation processing on the path evaluation data and the network preference data, so as to The best candidate path of the original network path is determined among the candidate network paths.
根据本公开实施例的第三个方面，提供一种电子设备，包括：处理器和存储器；其中，存储器上存储有计算机可读指令，所述计算机可读指令被所述处理器执行时实现上述任意示例性实施例中的网络路径的处理方法。According to a third aspect of the embodiments of the present disclosure, there is provided an electronic device, including: a processor and a memory; wherein, the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the above-mentioned A method for processing network paths in any exemplary embodiment.
根据本公开实施例的第四个方面，提供一种计算机可读存储介质，计算机可读存储介质上存储有计算机程序，所述计算机程序被处理器执行时实现上述任意示例性实施例中的网络路径的处理方法。According to a fourth aspect of the embodiments of the present disclosure, there is provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the network in any of the above-mentioned exemplary embodiments is implemented. The processing method for the path.
通过以下参照附图对本公开的示例性实施例的详细描述，本公开的其它特征及其优点将会变得清楚。Other features of the present disclosure and advantages thereof will become apparent through the following detailed description of exemplary embodiments of the present disclosure with reference to the accompanying drawings.
附图说明Description of drawings
此处的附图被并入说明书中并构成本说明书的一部分，示出了符合本公开的实施例，并与说明书一起用于解释本公开的原理。显而易见地，下面描述中的附图仅仅是本公开的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the disclosure and together with the description serve to explain the principles of the disclosure. Apparently, the drawings in the following description are only some embodiments of the present disclosure, and those skilled in the art can obtain other drawings according to these drawings without creative efforts.
图1示意性示出本公开一个实施例的网络路径的处理方法的流程示意图；FIG. 1 schematically shows a schematic flowchart of a method for processing a network path according to an embodiment of the present disclosure;
图2示意性示出本公开一个实施例的获取原始流量数据和备选流量数据的方法的流程示意图；FIG. 2 schematically shows a schematic flowchart of a method for acquiring original traffic data and alternative traffic data according to an embodiment of the present disclosure;
图3示意性示出本公开一个实施例的流量叠加处理的方法的流程示意图；FIG. 3 schematically shows a schematic flowchart of a method for traffic superposition processing according to an embodiment of the present disclosure;
图4示意性示出本公开一个实施例的路径评估处理的方法的流程示意图；FIG. 4 schematically shows a schematic flowchart of a method for path evaluation processing according to an embodiment of the present disclosure;
图5示意性示出本公开一个实施例的路径推荐处理的方法的流程示意图；FIG. 5 schematically shows a schematic flow chart of a method for path recommendation processing according to an embodiment of the present disclosure;
图6示意性示出本公开另一个实施例的路径推荐处理的方法的流程示意图；FIG. 6 schematically shows a schematic flowchart of a method for processing route recommendation according to another embodiment of the present disclosure;
图7示意性示出本公开一个实施例的应用场景下网络路径的处理方法的***架构图；FIG. 7 schematically shows a system architecture diagram of a network path processing method in an application scenario according to an embodiment of the present disclosure;
图8示意性示出本公开一个实施例的网络流量仿真模块的处理方法的流程示意图；FIG. 8 schematically shows a schematic flowchart of a processing method of a network traffic simulation module according to an embodiment of the present disclosure;
图9示意性示出本公开一个实施例的备选路径评估模块预测目标预测数据的方法的流程示意图；FIG. 9 schematically shows a schematic flowchart of a method for predicting target prediction data by an alternative path evaluation module according to an embodiment of the present disclosure;
图10示意性示出本公开一个实施例的备选路径推荐模块的处理方法的流程示意图；FIG. 10 schematically shows a schematic flowchart of a processing method of an alternative path recommendation module according to an embodiment of the present disclosure;
图11示意性示出本公开一个实施例的网络路径的处理装置的结构示意图；FIG. 11 schematically shows a schematic structural diagram of a network path processing device according to an embodiment of the present disclosure;
图12示意性示出本公开一个实施例的用于实现网络路径的处理方法的电子设备；FIG. 12 schematically shows an electronic device for implementing a network path processing method according to an embodiment of the present disclosure;
图13示意性示出本公开一个实施例的用于实现网络路径的处理方法的计算机可读存储介质。Fig. 13 schematically shows a computer-readable storage medium for implementing a network path processing method according to an embodiment of the present disclosure.
具体实施方式Detailed ways
现在将参考附图更全面地描述示例实施方式。然而，示例实施方式能够以多种形式实施，且不应被理解为限于在此阐述的范例；相反，提供这些实施方式使得本公开将更加全面和完整，并将示例实施方式的构思全面地传达给本领域的技术人员。所描述的特征、结构或特性可以以任何合适的方式结合在一个或更多实施方式中。在下面的描述中，提供许多具体细节从而给出对本公开的实施方式的充分理解。然而，本领域技术人员将意识到，可以实践本公开的技术方案而省略所述特定细节中的一个或更多，或者可以采用其它的方法、组元、装置、步骤等。在其它情况下，不详细示出或描述公知技术方案以避免喧宾夺主而使得本公开的各方面变得模糊。Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments may, however, be embodied in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete and will fully convey the concept of example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided in order to give a thorough understanding of embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced without one or more of the specific details being omitted, or other methods, components, devices, steps, etc. may be adopted. In other instances, well-known technical solutions have not been shown or described in detail to avoid obscuring aspects of the present disclosure.
本说明书中使用用语“一个”、“一”、“该”和“所述”用以表示存在一个或多个要素/组成部分/等；用语“包括”和“具有”用以表示开放式的包括在内的意思并且是指除了列出的要素/组成部分/等之外还可存在另外的要素/组成部分/等；用语“第一”和“第二”等仅作为标记使用，不是对其对象的数量限制。The terms "a", "an", "the" and "the" are used in this specification to indicate the existence of one or more elements/components/etc.; the terms "comprising" and "having" are used to indicate an open Included means and means that there may be additional elements/components/etc. in addition to the listed elements/components/etc; the terms "first" and "second" etc. The number of its objects is limited.
此外，附图仅为本公开的示意性图解，并非一定是按比例绘制。图中相同的附图标记表示相同或类似的部分，因而将省略对它们的重复描述。附图中所示的一些方框图是功能实体，不一定必须与物理或逻辑上独立的实体相对应。Furthermore, the drawings are merely schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically separate entities.
发明人注意到，相关技术中的路径选取方法只能基于一项性能指标进行选取，性能指标的综合考虑不足，导致路径选取的准确度低且效率差。并且，无法根据不同情况适应性选择路径，灵活性与适应性也不够。The inventor noticed that the path selection method in the related art can only be selected based on one performance index, and the comprehensive consideration of the performance index is insufficient, resulting in low accuracy and poor efficiency of path selection. Moreover, it is impossible to adaptively select paths according to different situations, and the flexibility and adaptability are not enough.
据此，本公开提供一种网络路径处理方案，能够更加精确地进行最优备选路径的选取，此外综合考虑多项影响路径选取的性能指标，从多个角度提升网络运维效率，同时满足用户在不同情况下的路径选择需求，适应性更强，灵活性更佳。Accordingly, the present disclosure provides a network path processing solution, which can more accurately select the optimal alternative path, and comprehensively consider multiple performance indicators that affect path selection, improve network operation and maintenance efficiency from multiple perspectives, and meet the requirements of The path selection needs of users in different situations are more adaptable and flexible.
图1示出了本公开一个实施例的网络路径的处理方法的流程图。在一些实施例中，下列的网络路径的处理方法由网络路径的处理装置执行。如图1所示，网络路径的处理方法至少包括以下步骤：Fig. 1 shows a flow chart of a method for processing a network path according to an embodiment of the present disclosure. In some embodiments, the following network path processing method is executed by a network path processing device. As shown in Figure 1, the method for processing a network path at least includes the following steps:
步骤S110.获取原始网络路径的原始流量数据和备选网络路径的备选流量数据， 对原始流量数据和备选流量数据进行流量叠加处理得到叠加流量数据。Step S110. Obtain the original traffic data of the original network path and the candidate traffic data of the candidate network path, and perform traffic superposition processing on the original traffic data and the candidate traffic data to obtain superimposed traffic data.
步骤S120.获取备选网络路径的网络路径数据，并对叠加流量数据和网络路径数据进行路径评估处理得到路径评估数据。Step S120. Obtain network path data of the candidate network path, and perform path evaluation processing on the superimposed traffic data and network path data to obtain path evaluation data.
步骤S130.获取与路径评估数据对应的网络偏好数据，对路径评估数据和网络偏好数据进行路径推荐处理，以在备选网络路径中确定原始网络路径的最优备选路径。Step S130. Acquire network preference data corresponding to the path evaluation data, and perform path recommendation processing on the path evaluation data and network preference data, so as to determine the optimal candidate path of the original network path among the candidate network paths.
在本公开的示例性实施例中，一方面，对原始流量数据和备选流量数据进行流量叠加处理，考虑了流量叠加过程中原始流量数据与备选流量数据不同的情况，能够更加精确地进行最优备选路径的选取，提升了确定的最优备选路径的准确度；另一方面，利用网络偏好数据和路径评估数据进行路径推荐处理，综合考虑多项影响路径选取的性能指标，从多个角度提升网络运维效率，同时满足用户在不同情况下的路径选择需求，适应性更强，灵活性更佳。In an exemplary embodiment of the present disclosure, on the one hand, the traffic superposition processing is performed on the original traffic data and the alternative traffic data, and the difference between the original traffic data and the alternative traffic data in the traffic superposition process can be considered, and more accurate The selection of the optimal alternative path improves the accuracy of the determined optimal alternative path; on the other hand, the network preference data and path evaluation data are used for path recommendation processing, and multiple performance indicators that affect path selection are comprehensively considered. Improve the efficiency of network operation and maintenance from multiple angles, and at the same time meet the path selection needs of users in different situations, with stronger adaptability and better flexibility.
下面对网络路径的处理方法的各个步骤进行详细说明。Each step of the network path processing method will be described in detail below.
在步骤S110中，获取原始网络路径的原始流量数据和备选网络路径的备选流量数据，对原始流量数据和备选流量数据进行流量叠加处理得到叠加流量数据。In step S110, the original traffic data of the original network path and the candidate traffic data of the candidate network path are acquired, and traffic superposition processing is performed on the original traffic data and the candidate traffic data to obtain superimposed traffic data.
在本公开的示例性实施例中，原始网络路径是在三层网络中，业务流在运行过程中正在使用的往返路径。而备选网络路径是在原始网络路径出现电路质差或者中断的情况下，能够重新选择的线路。In an exemplary embodiment of the present disclosure, the original network path is a round-trip path used by a service flow during operation in a three-layer network. The alternative network path is a line that can be reselected when the original network path has poor circuit quality or is interrupted.
在一些实施例中，图2示出了获取原始流量数据和备选流量数据的方法的流程示意图，如图2所示，该方法至少包括以下步骤：In some embodiments, FIG. 2 shows a schematic flowchart of a method for acquiring original traffic data and alternative traffic data. As shown in FIG. 2 , the method at least includes the following steps:
在步骤S210中，获取原始网络路径的原始路径数据和备选网络路径的备选路径数据，并对原始路径数据进行数据清洗处理得到原始清洗数据。In step S210, the original route data of the original network route and the candidate route data of the candidate network route are obtained, and data cleaning is performed on the original route data to obtain the original cleaned data.
在一些实施例中，原始路径数据中可以包括原始网络路径经过的各网络路由器端口的发送报文数量、接收报文数量、上行流量、下行流量、带宽、报文类型和业务的SLA性能评估数据，该SLA性能评估数据包括最大丢包率、最大单向时延和最大带宽利用率。除此之外，原始路径数据也可以根据实际情况和需求包括其他数据，本示例性实施例对此不做特殊限定。In some embodiments, the original path data may include the number of packets sent, the number of packets received, uplink traffic, downlink traffic, bandwidth, packet types, and SLA performance evaluation data of each network router port that the original network path passes through. , the SLA performance evaluation data includes maximum packet loss rate, maximum one-way delay, and maximum bandwidth utilization. In addition, the original route data may also include other data according to actual conditions and requirements, which is not specifically limited in this exemplary embodiment.
对应的，备选路径数据可以包括备选网络路径经过的各网络路由器端口的发送报文数量、接收报文数量、上行流量、下行流量、带宽、报文类型和业务的SLA性能评估数据，该SLA性能评估数据包括最大丢包率、最大单向时延和最大带宽利用率。除此之外，备选路径数据也可以根据实际情况和需求包括其他数据，本示例性实施例对 此不做特殊限定。Correspondingly, the candidate path data may include the number of packets sent, the number of packets received, uplink traffic, downlink traffic, bandwidth, packet types, and SLA performance evaluation data of each network router port that the alternative network path passes through. SLA performance evaluation data includes maximum packet loss rate, maximum one-way delay, and maximum bandwidth utilization. In addition, the alternative path data may also include other data according to actual conditions and requirements, which is not specifically limited in this exemplary embodiment.
在一些实施例中，可以对获得的原始路径数据进行数据清洗处理得到对应的原始清洗数据。In some embodiments, data cleaning may be performed on the obtained original path data to obtain corresponding original cleaning data.
例如，数据清洗处理可以包括各类性能数据的单位统一、缺失值处理和异常值处理等，也可以包括其他处理方式，本示例性实施例对此不做特殊限定。For example, the data cleaning process may include unit unification of various performance data, missing value processing, outlier value processing, etc., and may also include other processing methods, which are not specifically limited in this exemplary embodiment.
在步骤S220中，对原始清洗数据进行数据特征计算得到原始数据特征，并对原始数据特征和原始路径数据进行比较得到数据比较结果。In step S220, data feature calculation is performed on the original cleaning data to obtain the original data feature, and the original data feature is compared with the original path data to obtain a data comparison result.
对于数据清洗处理后的原始清洗数据，还可以进行小时的流量峰值计算，以统一按照小时粒度的接收报文数量和发送报文数量等。For the original cleaned data after data cleaning, hourly traffic peak calculations can also be performed to unify the number of received packets and the number of sent packets at an hourly granularity.
在一些实施例中，还可以对原始清洗数据进行数据特征计算得到原始数据特征。In some embodiments, data feature calculation may also be performed on the original cleaned data to obtain original data features.
例如，该原始数据特征可以包括单向时延、双向时延、丢包率和带宽利用率。For example, the raw data characteristics may include one-way delay, two-way delay, packet loss rate, and bandwidth utilization rate.
例如，单向时延＝Egress(出口)节点报文接收时间-Ingress(入口)节点报文发送时间；双向时延＝(Egress节点报文接收时间-Ingress节点报文发送时间)+Ingress节点报文接收时间-Egress节点报文发送时间)；丢包率＝(Egress节点接收报文数-Ingress节点发送报文数)*100％/Ingress节点发送报文数；带宽利用率＝流量/带宽。For example, one-way delay=Egress (exit) node message receiving time-Ingress (entry) node message sending time; two-way time delay=(Egress node message receiving time-Ingress node message sending time)+Ingress node report Time for receiving text-Egress node message sending time); packet loss rate=(number of messages received by Egress node-number of messages sent by Ingress node)*100%/number of messages sent by Ingress node; bandwidth utilization rate=traffic/bandwidth.
在一些实施例中，可以对原始数据特征和原始路径数据进行比较。In some embodiments, raw data features and raw route data may be compared.
例如，将单向时延与原始路径数据中SLA性能评估数据的最大单向时延进行比较；将丢包率与原始路径数据中SLA性能评估数据的最大丢包率进行比较；将带宽利用率与原始路径数据中SLA性能评估数据的最大带宽利用率进行比较。For example, compare the one-way delay with the maximum one-way delay of the SLA performance evaluation data in the original path data; compare the packet loss rate with the maximum packet loss rate of the SLA performance evaluation data in the original path data; compare the bandwidth utilization Compare with the maximum bandwidth utilization of the SLA performance evaluation data in the original path data.
除此之外，当SLA性能评估数据中包括最大双向时延时，还可以将双向时延与该最大双向时延进行比较。In addition, when the SLA performance evaluation data includes the maximum two-way delay, the two-way delay may also be compared with the maximum two-way delay.
在一些实施例中，根据该比较方式可以得到数据比较结果。In some embodiments, the data comparison result can be obtained according to the comparison manner.
在步骤S230中，若数据比较结果为原始数据特征未满足原始路径数据中包括的服务级别协议SLA性能评估数据的要求，确定原始网络路径异常，并获取原始路径数据中的原始流量数据和备选路径数据中的备选流量数据。In step S230, if the result of the data comparison is that the characteristics of the original data do not meet the requirements of the service level agreement (SLA) performance evaluation data included in the original path data, it is determined that the original network path is abnormal, and the original traffic data and alternatives in the original path data are obtained. Alternative traffic data in path data.
例如，当丢包率为100％时，可以确定原始网络路径出现中断情况，因此原始网络路径异常。当丢包率不为100％，但是其他原始数据特征超过原始路径数据时，确定原始网络路径出现质差情况，因此原始网络路径异常。For example, when the packet loss rate is 100%, it can be determined that the original network path is interrupted, so the original network path is abnormal. When the packet loss rate is not 100%, but other original data characteristics exceed the original path data, it is determined that the original network path has poor quality, and therefore the original network path is abnormal.
除此之外，原始网络路径的中断情况也可以是采集到用户投诉的中断信息确定的， 本示例性实施例对此不做特殊限定。In addition, the interruption of the original network path may also be determined by collecting the interruption information complained by the user, which is not specifically limited in this exemplary embodiment.
值得说明的是，也可以不把原始路径数据作为原始数据特征比较的门限值，该门限值还可以是根据历史情况设定的，例如带宽利用率门限值为80％，时延10ms等，本示例性实施例对此不做特殊限定。It is worth noting that the original path data may not be used as the threshold value for feature comparison of the original data, and the threshold value may also be set according to historical conditions, for example, the bandwidth utilization threshold value is 80%, and the delay is 10ms etc., which is not specifically limited in this exemplary embodiment.
在一些实施例中，当确定原始网络路径异常时，可以获取原始路径数据中的原始流量数据和备选路径数据中的备选流量数据。原始流量数据包括原始路径数据中的上行流量和下行流量，备选流量数据包括备选路径数据中的上行流量和下行流量。In some embodiments, when it is determined that the original network path is abnormal, the original flow data in the original path data and the candidate flow data in the candidate path data may be acquired. The original traffic data includes uplink traffic and downlink traffic in the original path data, and the candidate traffic data includes uplink traffic and downlink traffic in the candidate path data.
当没有原始数据特征超过限定值时，可以确定原始网络路径正常，无需获取原始路径数据中的原始流量数据和备选路径数据中的备选流量数据进行后续处理。When no characteristic of the original data exceeds the limit value, it can be determined that the original network path is normal, and there is no need to obtain the original flow data in the original path data and the candidate flow data in the candidate path data for subsequent processing.
在本示例性实施例中，通过原始路径数据和备选路径数据确定原始网络路径的异常情况，以获取原始流量数据和备选流量数据为最优备选路径的确定提供了前置逻辑判断，并且为后续处理过程提供了理论基础和数据基础。In this exemplary embodiment, the abnormality of the original network path is determined through the original path data and the alternative path data, so as to obtain the original traffic data and the alternative traffic data to provide a pre-logic judgment for determining the optimal alternative path, And it provides the theoretical basis and data basis for the subsequent processing.
在获取到原始流量数据和备选流量数据之后，可以对原始流量数据和备选流量数据进行流量叠加处理。After the original traffic data and the candidate traffic data are acquired, traffic superposition processing may be performed on the original traffic data and the candidate traffic data.
在一些实施例中，图3示出了流量叠加处理的方法的流程示意图，如图3所示，该方法至少包括以下步骤：In some embodiments, FIG. 3 shows a schematic flowchart of a method for traffic superposition processing. As shown in FIG. 3 , the method at least includes the following steps:
在步骤S310中，对原始路径数据中的原始类型数据和备选路径数据中的备选类型数据进行报文类型比较得到类型比较结果。In step S310, the packet type comparison is performed on the original type data in the original route data and the candidate type data in the candidate route data to obtain a type comparison result.
获取原始路径数据中的报文类型作为原始类型数据，并获取备选路径数据中的报文类型作为备选类型数据。在一些实施例中，可以判断原始类型数据与备选类型数据是否相同，以得到类型比较结果。The packet type in the original path data is obtained as the original type data, and the packet type in the alternate path data is obtained as the alternate type data. In some embodiments, it may be determined whether the original type data is the same as the alternative type data, so as to obtain a type comparison result.
在步骤S320中，若类型比较结果为原始类型数据与备选类型数据相同，对原始流量数据和备选流量数据进行流量叠加处理得到叠加流量数据。In step S320, if the result of type comparison is that the original type data is the same as the alternative type data, traffic superimposition processing is performed on the original traffic data and the alternative traffic data to obtain superimposed traffic data.
当原始类型数据与备选类型数据相同时，表明原始路径数据的上行流量和下行流量，以及备选路径数据的上行流量和下行流量是相同的，因此，可以对原始流量数据和备选流量数据进行流量叠加处理得到叠加流量数据。When the original type data is the same as the alternate type data, it indicates that the upstream traffic and downstream traffic of the original path data and the upstream traffic and downstream traffic of the alternate path data are the same, therefore, the original traffic data and the alternate traffic data can be compared Perform traffic superposition processing to obtain superimposed traffic data.
值得说明的是，流量叠加处理时，可以是原始路径数据中的上行流量与备选路径数据的上行流量进行叠加，原始路径数据的下行流量和备选路径数据中的下行流量进行叠加。因此，叠加流量数据可以包括流量叠加后的上行流量和叠加后的下行流量。It is worth noting that, during traffic superposition processing, the uplink traffic in the original path data and the uplink traffic in the alternative path data may be superimposed, and the downlink traffic in the original path data and the downlink traffic in the alternative path data may be superimposed. Therefore, the superimposed traffic data may include superimposed uplink traffic and superimposed downlink traffic.
在步骤S330中，若类型比较结果为原始类型数据与备选类型数据不同，对原始路 径数据和备选路径数据进行叠加更新计算得到叠加流量数据。In step S330, if the result of the type comparison is that the original type data is different from the alternative type data, the original route data and the alternative route data are superimposed and updated to obtain superimposed traffic data.
当原始类型数据与备选类型数据不同时，表明原始路径数据的上行流量和下行流量，以及备选路径数据的上行流量和下行流量是不同的，因此，可以对原始流量数据和备选流量数据进行流量更新计算得到叠加流量数据。When the original type data is different from the alternative type data, it indicates that the uplink traffic and downlink traffic of the original path data and the uplink traffic and downlink traffic of the alternative path data are different, therefore, the original traffic data and the alternative traffic data can be compared Perform flow update calculations to obtain superimposed flow data.
例如，新业务流量＝新业务报文数量×备选路径报文长度，而该备选路径报文长度＝备选路径流量/备选路径报文数量。For example, the new service flow=the number of new service packets×the packet length of the alternative path, and the packet length of the alternative path=the traffic volume of the alternative path/the number of packets of the alternative path.
由于是将原始网络路径的上行流量和下行流量叠加至备选网络路径的上行流量和下行流量中，因此，该新业务流量即为叠加后的叠加流量数据。Since the uplink traffic and downlink traffic of the original network path are superimposed into the uplink traffic and downlink traffic of the alternative network path, the new service traffic is superimposed superimposed traffic data.
在计算备选路径报文长度时所使用的备选路径流量可以是备选网络路径的上行流量或下行流量，所使用的备选路径报文数量可以是备选网络路径的发送报文数量或接收报文数量，本示例性实施例对此不做特殊限定。The alternative path traffic used when calculating the length of the alternative path message can be the upstream or downstream traffic of the alternative network path, and the number of alternative path packets used can be the number of packets sent by the alternative network path or The number of received packets is not specifically limited in this exemplary embodiment.
在本示例性实施例中，根据原始类型数据和备选类型数据的不同类型比较结果可以对应有得到叠加流量数据的不同方式，对不同报文类型情况的处理逻辑完整严谨，保证了叠加流量数据的准确性。In this exemplary embodiment, according to the comparison results of different types of original type data and alternative type data, there are different ways to obtain superimposed traffic data, and the processing logic for different message types is complete and rigorous, ensuring superimposed traffic data accuracy.
在步骤S120中，获取备选网络路径的网络路径数据，并对叠加流量数据和网络路径数据进行路径评估处理得到路径评估数据。In step S120, network path data of the candidate network path is obtained, and path evaluation processing is performed on the superimposed traffic data and network path data to obtain path evaluation data.
在本公开的示例性实施例中，在得到叠加流量数据之后，可以进一步对叠加流量数据进行路径评估处理。In an exemplary embodiment of the present disclosure, after the superimposed traffic data is obtained, path evaluation processing may be further performed on the superimposed traffic data.
在一些实施例中，图4示出了路径评估处理的方法的流程示意图，如图4所示，该方法至少包括以下步骤：In some embodiments, FIG. 4 shows a schematic flowchart of a method for path evaluation processing. As shown in FIG. 4, the method at least includes the following steps:
在步骤S410中，获取备选路径数据中的网络路径数据，并对叠加流量数据和网络路径数据进行路径评估处理得到网络评估数据。In step S410, network path data in the candidate path data is obtained, and path evaluation processing is performed on the superimposed traffic data and network path data to obtain network evaluation data.
为进行路径评估处理，可以获取到的备选路径数据中的网络路径数据可以包括备选网络路径中的路由个数、带宽、链路光纤总长度，也可以包括其他能够进行路径评估处理的数据，本示例性实施例对此不做特殊限定。For path evaluation processing, the network path data in the candidate path data that can be obtained may include the number of routes in the candidate network path, the bandwidth, the total length of link fibers, and other data that can be processed for path evaluation , which is not specifically limited in this exemplary embodiment.
在一些实施例中，可以对叠加流量数据和网络路径数据进行路径评估处理。In some embodiments, path evaluation processing may be performed on overlay traffic data and network path data.
例如，For example,
路由跳数＝路径中路由个数-1Routing hops = number of routes in the path - 1
带宽利用率＝备选路径总流量/备选路径带宽        (1)Bandwidth utilization = total traffic of alternative paths / bandwidth of alternative paths (1)
备选路径长度＝备选路径的链路光纤总长度。该路由跳数、带宽利用率和备选路径 长度即为网络评估数据。Alternative path length = the total length of the link fiber of the alternate path. The routing hops, bandwidth utilization and alternative path lengths are the network evaluation data.
在步骤S420中，获取当前时间戳和备选网络路径的历史预测数据，并利用当前时间戳和历史预测数据进行时延预测处理得到目标预测数据，以确定网络评估数据和目标预测数据为路径评估数据。In step S420, the current time stamp and the historical forecast data of the alternative network path are obtained, and the delay prediction process is performed using the current time stamp and the historical forecast data to obtain the target forecast data, so as to determine the network evaluation data and the target forecast data as path evaluation data.
在一些实施例中，该历史预测数据可以包括历史流量数据和丢包率。例如，历史流量数据可以是以当前时间为节点，汇总的当前时间之前的所有流量数据。In some embodiments, the historical forecast data may include historical traffic data and packet loss rate. For example, the historical traffic data may take the current time as a node and summarize all the traffic data before the current time.
在一些实施例中，可以利用当前时间戳和历史预测数据进行时延预测处理。In some embodiments, the latency prediction process can be performed using the current timestamp and historical prediction data.
在一些实施例中，将当前时间戳和历史预测数据输入至预先训练好的时延预测模型中，以使时延预测模型输出目标预测数据。In some embodiments, the current time stamp and historical prediction data are input into a pre-trained delay prediction model, so that the delay prediction model outputs target prediction data.
目标预测数据包括单向时延数据和预测的丢包率。单向时延数据即为收发业务流时间差。该收发业务流时间差＝目的端接收业务报文时间-源端发送业务报文时间。The target prediction data includes one-way delay data and predicted packet loss rate. The one-way delay data is the time difference between sending and receiving service flows. The time difference between sending and receiving service flows = the time when the destination end receives the service message - the time when the source end sends the service message.
举例而言，该时延预测模型可以是LightGBM模型。For example, the delay prediction model may be a LightGBM model.
LightGBM算法是一个实现GBDT(Gradient Boosting Decision Tree，梯度提升决策树)思想的机器学习算法，预测算法公式为：The LightGBM algorithm is a machine learning algorithm that implements the idea of GBDT (Gradient Boosting Decision Tree, gradient boosting decision tree). The prediction algorithm formula is:
f m(x)＝f m-1(x)+T(x；θ m)                  (2) f m (x) = f m-1 (x) + T (x; θ m ) (2)
其中，x为预测样本，T(x；θ m)表示决策树，θ m表示决策树参数，m为树的个数，f m(x)为样本预测值。 Among them, x is the predicted sample, T(x; θ m ) represents the decision tree, θ m represents the parameters of the decision tree, m is the number of trees, and f m (x) is the predicted value of the sample.
损失函数表示为：The loss function is expressed as:
Figure PCTCN2022112000-appb-000001
Figure PCTCN2022112000-appb-000001
其中，y i为第i个样本的真实值，f m(x i)为第i个样本的预测值。 Among them, y i is the true value of the i-th sample, and f m ( xi ) is the predicted value of the i-th sample.
根据公式(1)极小化损失函数得到参数θ m： Minimize the loss function according to formula (1) to get the parameter θ m :
Figure PCTCN2022112000-appb-000002
Figure PCTCN2022112000-appb-000002
在一些实施例中，可以对LightGBM模型进行训练。In some embodiments, the LightGBM model can be trained.
例如，获取备选网络路径的历史流量样本、时间戳样本、单向时延样本和丢包率样本，并将样本数据按照7：3的比例分为训练集和测试集。例如，基于训练集训练LightGBM模型，并基于测试集对LightGBM模型的参数进行调整。For example, obtain historical traffic samples, time stamp samples, one-way delay samples, and packet loss rate samples of alternative network paths, and divide the sample data into a training set and a test set at a ratio of 7:3. For example, the LightGBM model is trained based on the training set, and the parameters of the LightGBM model are adjusted based on the test set.
首先，输入历史流量样本、时间戳样本、单向时延样本和丢包率样本，并初始化f 0(x i)；然后，拟合残差树T(x；θ m)，并基于公式(3)和公式(4)更新公式(2)，以此类推，不断重复拟合步骤和更新步骤进行迭代直至收敛，以得到训练好的LightGBM模型。 First, input historical traffic samples, time stamp samples, one-way delay samples and packet loss rate samples, and initialize f 0 (xi ) ; then, fit the residual tree T(x; θ m ), and based on the formula ( 3) and formula (4) to update formula (2), and so on, repeating the fitting step and updating step until convergence, so as to obtain the trained LightGBM model.
在训练好LightGBM模型之后，可以将当前时间戳和历史预测数据输入至已训练好的LightGBM模型中进行目标预测数据的预测，因此，该LightGBM模型可以输出预测出的单向时延数据和丢包率作为目标预测数据。After the LightGBM model is trained, the current timestamp and historical forecast data can be input into the trained LightGBM model to predict the target forecast data. Therefore, the LightGBM model can output the predicted one-way delay data and packet loss rate as the target forecast data.
在得到目标预测数据之后，可以确定该目标预测数据和网络评估数据为路径评估数据。After the target prediction data is obtained, the target prediction data and network evaluation data can be determined as path evaluation data.
在本示例性实施例中，对叠加流量数据和网络路径数据进行路径评估处理可以得到路径评估数据，为确定最优备选路径提供了数据基础，使得最优备选路径的确定可以综合多项性能，保证了最优备选路径的确定效果和网络运维效率。In this exemplary embodiment, path evaluation data can be obtained by performing path evaluation processing on superimposed traffic data and network path data, which provides a data basis for determining the optimal alternative path, so that the determination of the optimal alternative path can synthesize multiple performance, which ensures the determination of the optimal alternative path and the efficiency of network operation and maintenance.
在步骤S130中，获取与路径评估数据对应的网络偏好数据，对路径评估数据和网络偏好数据进行路径推荐处理，以在备选网络路径中确定原始网络路径的最优备选路径。In step S130, network preference data corresponding to the path evaluation data is obtained, and path recommendation processing is performed on the path evaluation data and network preference data, so as to determine the optimal candidate path of the original network path among the candidate network paths.
在本公开的示例性实施例中，为使最优备选路径的确定可以综合考虑用户喜好，因此还可以进一步获取网络偏好数据。In an exemplary embodiment of the present disclosure, in order to determine the optimal alternative path, user preferences may be considered comprehensively, so network preference data may be further acquired.
该网络偏好数据是用户对路径评估数据中路由跳数、带宽利用率、备选路径长度和单向时延数据的偏好数据。举例而言，在n个用户中，30％偏好短路径、40％偏好少跳数、10％偏好短时延、10％偏好少带宽利用率、10％偏好丢包率，因此，可以得到网络偏好数据w 1＝-0.3，w 2＝-0.4，w 3＝-0.1，w 4＝-0.1，w 5＝-0.1。 The network preference data is the user's preference data on the route hop count, bandwidth utilization rate, candidate path length and one-way delay data in the path evaluation data. For example, among n users, 30% prefer short paths, 40% prefer less hops, 10% prefer short delays, 10% prefer less bandwidth utilization, and 10% prefer packet loss rates. Therefore, the network can be obtained Preference data w 1 =-0.3, w 2 =-0.4, w 3 =-0.1, w 4 =-0.1, w 5 =-0.1.
在得到网络偏好数据之后，可以对路径评估数据和网络偏好数据进行路径推荐处理。After obtaining the network preference data, path recommendation processing can be performed on the path evaluation data and the network preference data.
在一些实施例中，图5示出了路径推荐处理的方法的流程示意图，如图5所示，该方法至少包括以下步骤：In some embodiments, FIG. 5 shows a schematic flowchart of a method for route recommendation processing. As shown in FIG. 5, the method at least includes the following steps:
在步骤S510中，对备选路径数据进行数据清洗处理得到备选清洗数据，并对备选清洗数据进行数据特征计算得到备选数据特征。In step S510, data cleaning processing is performed on the candidate path data to obtain candidate cleaning data, and data feature calculation is performed on the candidate cleaning data to obtain candidate data features.
在一些实施例中，数据清洗处理可以包括备选路径数据的单位统一、缺失值处理和异常值处理等，也可以包括其他处理方式，本示例性实施例对此不做特殊限定。In some embodiments, the data cleaning process may include unit unification of candidate path data, missing value processing, outlier value processing, etc., and may also include other processing methods, which are not specifically limited in this exemplary embodiment.
对于数据清洗处理后的备选清洗数据，还可以进行小时的流量峰值计算，以统一按照小时粒度的接收报文数量和发送报文数量等。For the alternative cleaning data after data cleaning processing, hourly traffic peak calculations can also be performed to unify the number of received packets and the number of sent packets at an hourly granularity.
在一些实施例中，还可以对备选清洗数据进行数据特征计算得到备选数据特征。In some embodiments, data feature calculation may also be performed on the candidate cleaning data to obtain candidate data features.
在一些实施例中，该备选数据特征可以包括单向时延、双向时延、丢包率和带宽利用率。In some embodiments, the candidate data characteristics may include one-way delay, two-way delay, packet loss rate, and bandwidth utilization.
例如，单向时延＝Egress节点报文接收时间-Ingress节点报文发送时间；双向时延＝(Egress节点报文接收时间-Ingress节点报文发送时间)+Ingress节点报文接收时间-Egress节点报文发送时间)；丢包率＝(Egress节点接收报文数-Ingress节点发送报文数)*100％/Ingress节点发送报文数；带宽利用率＝流量/带宽。For example, one-way delay=Egress node message receiving time-Ingress node message sending time; two-way delay=(Egress node message receiving time-Ingress node message sending time)+Ingress node message receiving time-Egress node message sending time); packet loss rate=(the number of messages received by the Egress node-the number of messages sent by the Ingress node)*100%/the number of messages sent by the Ingress node; bandwidth utilization rate=traffic/bandwidth.
在步骤S520中，获取备选清洗数据中的特征门限数据，并对备选特征数据和特征门限数据进行比较到备选比较结果。In step S520, the characteristic threshold data in the candidate cleaning data is obtained, and the candidate characteristic data and the characteristic threshold data are compared to a candidate comparison result.
从备选清洗数据中获取到的特征门限数据可以包括备选网络路径的SLA性能评估数据，亦即最大丢包率、最大单向时延和最大带宽利用率。The feature threshold data obtained from the candidate cleaning data may include the SLA performance evaluation data of the candidate network path, that is, the maximum packet loss rate, the maximum one-way delay and the maximum bandwidth utilization.
在一些实施例中，可以对备选特征数据和特征门限数据进行比较。In some embodiments, the candidate feature data can be compared to feature threshold data.
例如，可以是对备选特征数据和特征门限数据进行除法计算得到二者比值，以作为备选比较结果。For example, a ratio between the candidate characteristic data and the characteristic threshold data may be calculated to obtain the ratio of the two as the candidate comparison result.
在步骤S530中，基于备选比较结果，对路径评估数据和网络偏好数据进行路径推荐处理，以在备选网络路径中确定原始网络路径的最优备选路径。In step S530, based on the alternative comparison results, path recommendation processing is performed on the path evaluation data and network preference data, so as to determine the optimal alternative path of the original network path among the alternative network paths.
当备选比较结果为备选特征数据和特征门限数据的比值小于或等于1时，可以继续对备选网络路径的路径评估数据和网络偏好数据进行路径推荐处理。当备选比较结果为备选特征数据和特征门限数据的比值大于1时，可以将该备选网络路径舍弃，对其他备选网络路径的路径评估数据和网络偏好数据进行路径推荐处理。When the candidate comparison result is that the ratio of the candidate feature data to the feature threshold data is less than or equal to 1, the path recommendation process may continue to be performed on the path evaluation data and network preference data of the candidate network path. When the candidate comparison result is that the ratio of the candidate feature data to the feature threshold data is greater than 1, the candidate network path can be discarded, and path recommendation processing is performed on the path evaluation data and network preference data of other candidate network paths.
在一些实施例中，图6示出了进一步进行路径推荐处理的方法的流程示意图，如图6所示，该方法至少包括以下步骤：In some embodiments, FIG. 6 shows a schematic flowchart of a method for further processing path recommendation. As shown in FIG. 6, the method at least includes the following steps:
在步骤S610中，对路径评估数据进行数据标准化得到标准评估数据，并对标准评估数据和网络偏好数据进行路径推荐处理得到与备选网络路径对应的路径推荐参数。In step S610, data standardization is performed on the path evaluation data to obtain standard evaluation data, and path recommendation processing is performed on the standard evaluation data and network preference data to obtain path recommendation parameters corresponding to candidate network paths.
在一些实施例中，可以利用Z-score算法对路径评估数据进行数据标准化：In some embodiments, the Z-score algorithm can be used to standardize the path evaluation data:
Figure PCTCN2022112000-appb-000003
Figure PCTCN2022112000-appb-000003
其中，μ为多条备选网络路径的路径评估数据，亦即路由跳数、带宽利用率、备选路径长度和单向时延数据的对应均值。σ为多条备选网络路径的路径评估数据的对应标准差。Among them, μ is the path evaluation data of multiple candidate network paths, that is, the corresponding average value of routing hops, bandwidth utilization, candidate path length and one-way delay data. σ is the corresponding standard deviation of the path evaluation data of multiple candidate network paths.
例如，当有5条备选网络路径，且计算备选路径长度时，μ为5条备选网络路径的备选路径长度均值，σ为5条备选网络路径的备选路径长度的标准差。以此类推，也可以计算另外三项的路径评估数据得到对应的标准评估数据。For example, when there are 5 candidate network paths and the length of the candidate paths is calculated, μ is the mean value of the candidate path lengths of the 5 candidate network paths, and σ is the standard deviation of the candidate path lengths of the 5 candidate network paths . By analogy, the path evaluation data of the other three items can also be calculated to obtain the corresponding standard evaluation data.
标准评估数据中备选路径长度、路由跳数、单向时延数据、带宽利用率和丢包率 的网络偏好数据分别为为w 1、w 2、w 3、w 4和w 5。并且，由于五项指标越小，路径最优，因此五项网络偏好数据都小于0。 The network preference data of the alternative path length, routing hops, one-way delay data, bandwidth utilization rate and packet loss rate in the standard evaluation data are w 1 , w 2 , w 3 , w 4 and w 5 respectively. Moreover, since the smaller the five indexes are, the path is optimal, so the five network preference data are all less than 0.
在一些实施例中，对标准评估数据和网络偏好数据进行路径推荐处理时，可以以网络偏好数据作为权重进行计算：In some embodiments, when performing path recommendation processing on the standard evaluation data and network preference data, the network preference data can be used as the weight for calculation:
Figure PCTCN2022112000-appb-000004
Figure PCTCN2022112000-appb-000004
其中，w 1+w 2+w 3+w 4+w 5＝-1，且w 1<0、w 2<0、w 3<0、w 4<0、w 5<0，μ 1为多条备选网络路径的备选路径长度的均值，σ 1为多条备选网络路径的备选路径长度的标准差，μ 2为多条备选网络路径的路由跳数的均值，σ 2为多条备选网络路径的路由跳数的标准差，μ 3为多条备选网络路径的单向时延数据的均值，σ 3为多条备选网络路径的单向时延数据的标准差，μ 4为多条备选网络路径的带宽利用率的均值，σ 4为多条备选网络路径的带宽利用率的标准差，μ 5为多条备选网络路径的丢包率的均值，σ 5为多条备选网络路径的丢包率的标准差。Y为路径推荐参数，该路径推荐参数的值越大，表示推荐程度越高。 Among them, w 1 +w 2 +w 3 +w 4 +w 5 =-1, and w 1 <0, w 2 <0, w 3 <0, w 4 <0, w 5 <0, μ 1 is more The mean value of the candidate path lengths of the candidate network paths, σ 1 is the standard deviation of the candidate path lengths of the multiple candidate network paths, μ 2 is the mean value of the routing hops of the multiple candidate network paths, and σ 2 is The standard deviation of the routing hops of multiple alternative network paths, μ 3 is the mean value of the one-way delay data of multiple alternative network paths, and σ 3 is the standard deviation of the one-way delay data of multiple alternative network paths , μ 4 is the average value of the bandwidth utilization ratio of multiple alternative network paths, σ 4 is the standard deviation of the bandwidth utilization ratio of multiple alternative network paths, μ 5 is the average value of the packet loss rate of multiple alternative network paths, σ 5 is the standard deviation of the packet loss rate of multiple alternative network paths. Y is a path recommendation parameter, and a larger value of the path recommendation parameter indicates a higher degree of recommendation.
在步骤S620中，对路径推荐参数进行推荐程度识别，以在备选网络路径中确定原始网络路径的最优备选路径。In step S620, the recommendation degree identification is performed on the path recommendation parameters, so as to determine the optimal candidate path of the original network path among the candidate network paths.
在得到路径推荐参数之后，可以对多条备选网络路径的路径推荐参数进行比较，以确定路径推荐参数最大的备选网络路径作为原始网络路径的最优备选路径。After the path recommendation parameters are obtained, the path recommendation parameters of multiple candidate network paths can be compared to determine the candidate network path with the largest path recommendation parameter as the optimal candidate path of the original network path.
在本示例性实施例中，结合用户的网络偏好数据和路径评估数据进行路径推荐处理确定最优备选路径，既能满足用户需求，还能进一步提升网络运维效率，推荐效果更佳。In this exemplary embodiment, combining the user's network preference data and path evaluation data to perform path recommendation processing to determine the optimal alternative path can not only meet user needs, but also further improve network operation and maintenance efficiency, and the recommendation effect is better.
下面结合一应用场景对本公开实施例中网络路径的处理方法做出详细说明。The method for processing a network path in an embodiment of the present disclosure will be described in detail below in conjunction with an application scenario.
图7示出了应用场景下网络路径的处理方法的***架构图，如图7所示，该***架构中包括客户设备、路由器、客户业务中心、数据采集模块、数据存储模块、数据处理模块、端到端业务质量评估模块、网络流量仿真模块、备选路径评估模块和备选路径推荐模块。Figure 7 shows a system architecture diagram of a network path processing method in an application scenario. As shown in Figure 7, the system architecture includes a client device, a router, a customer service center, a data collection module, a data storage module, a data processing module, An end-to-end service quality assessment module, a network flow simulation module, an alternative path evaluation module and an alternative path recommendation module.
其中，客户设备为客户侧边缘网络设备。Wherein, the client device is a client-side edge network device.
路由器为IP(Internet Protocol，网际互连协议)网络路由器设备。The router is an IP (Internet Protocol, Internet Interconnection Protocol) network router device.
客户业务中心为客户或运营商提供业务的云端服务器集群。The customer business center is a cloud server cluster that provides services for customers or operators.
数据采集模块可以采集原始网络路径的原始路径数据和备选网络路径的备选路径 数据。The data collection module can collect the original path data of the original network path and the candidate path data of the candidate network path.
原始路径数据中可以包括原始网络路径经过的各网络路由器端口的发送报文数量、接收报文数量、上行流量、下行流量、带宽、报文类型和业务的SLA性能评估数据，该SLA性能评估数据包括最大丢包率、最大单向时延和最大带宽利用率。除此之外，原始路径数据也可以根据实际情况和需求包括其他数据，本示例性实施例对此不做特殊限定。The original path data may include the number of packets sent, the number of packets received, upstream traffic, downstream traffic, bandwidth, packet types, and SLA performance evaluation data of each network router port that the original network path passes through. The SLA performance evaluation data Including maximum packet loss rate, maximum one-way delay and maximum bandwidth utilization. In addition, the original route data may also include other data according to actual conditions and requirements, which is not specifically limited in this exemplary embodiment.
对应的，备选路径数据可以包括备选网络路径经过的各网络路由器端口的发送报文数量、接收报文数量、上行流量、下行流量、带宽、报文类型和业务的SLA性能评估数据，该SLA性能评估数据包括最大丢包率、最大单向时延和最大带宽利用率。除此之外，备选路径数据也可以根据实际情况和需求包括其他数据，本示例性实施例对此不做特殊限定。Correspondingly, the candidate path data may include the number of packets sent, the number of packets received, uplink traffic, downlink traffic, bandwidth, packet types, and SLA performance evaluation data of each network router port that the alternative network path passes through. SLA performance evaluation data includes maximum packet loss rate, maximum one-way delay, and maximum bandwidth utilization. In addition, the candidate path data may also include other data according to actual conditions and requirements, which is not specifically limited in this exemplary embodiment.
数据存储模块可以高效快速的存储海量采集到的原始路径数据和备选路径数据。The data storage module can efficiently and quickly store a large amount of collected original path data and alternative path data.
在一些实施例中，可以对获得的原始路径数据和备选路径数据进行数据清洗处理得到对应的原始清洗数据和备选清洗数据。In some embodiments, data cleaning may be performed on the obtained original path data and candidate path data to obtain corresponding original cleaning data and candidate cleaning data.
例如，数据清洗处理可以包括各类性能数据的单位统一、缺失值处理和异常值处理等，也可以包括其他处理方式，本示例性实施例对此不做特殊限定。For example, the data cleaning process may include unit unification of various performance data, missing value processing, outlier value processing, etc., and may also include other processing methods, which are not specifically limited in this exemplary embodiment.
数据处理模块可以处理采集到的原始清洗数据和备选清洗数据。The data processing module can process the collected original cleaning data and alternative cleaning data.
对于数据清洗处理后的原始清洗数据和备选清洗数据，还可以进行小时的流量峰值计算，以统一按照小时粒度的接收报文数量和发送报文数量等。For the original cleaned data and alternative cleaned data after data cleaning, hourly traffic peak calculation can also be performed to unify the number of received packets and sent packets according to hourly granularity.
在一些实施例中，还可以对原始清洗数据进行数据特征计算得到原始数据特征，并对对备选清洗数据进行数据特征计算得到备选数据特征。In some embodiments, data feature calculation may be performed on the original cleaned data to obtain original data features, and data feature calculation may be performed on candidate cleaned data to obtain candidate data features.
在一些实施例中，该原始数据特征和备选数据特征中都可以包括单向时延、双向时延、丢包率和带宽利用率。In some embodiments, both the original data feature and the candidate data feature may include one-way delay, two-way delay, packet loss rate, and bandwidth utilization rate.
例如，单向时延＝Egress节点报文接收时间-Ingress节点报文发送时间；双向时延＝(Egress节点报文接收时间-Ingress节点报文发送时间)+Ingress节点报文接收时间-Egress节点报文发送时间)；丢包率＝(Egress节点接收报文数-Ingress节点发送报文数)*100％/Ingress节点发送报文数；带宽利用率＝流量/带宽。For example, one-way delay=Egress node message receiving time-Ingress node message sending time; two-way delay=(Egress node message receiving time-Ingress node message sending time)+Ingress node message receiving time-Egress node message sending time); packet loss rate=(the number of messages received by the Egress node-the number of messages sent by the Ingress node)*100%/the number of messages sent by the Ingress node; bandwidth utilization rate=traffic/bandwidth.
端到端业务质量评估模块运用端到端业务质量评估方法对端到端业务质量进行评估。The end-to-end service quality evaluation module uses the end-to-end service quality evaluation method to evaluate the end-to-end service quality.
在一些实施例中，可以对原始数据特征和原始路径数据进行比较评估出正常、质 差和中断三种数据比较结果。In some embodiments, the characteristics of the original data can be compared with the original path data to evaluate three data comparison results: normal, poor quality and interruption.
在一些实施例中，将单向时延与原始路径数据中SLA性能评估数据的最大单向时延进行比较；将丢包率与原始路径数据中SLA性能评估数据的最大丢包率进行比较；将带宽利用率与原始路径数据中SLA性能评估数据的最大带宽利用率进行比较。In some embodiments, comparing the one-way delay with the maximum one-way delay of the SLA performance evaluation data in the original path data; comparing the packet loss rate with the maximum packet loss rate of the SLA performance evaluation data in the original path data; Compare the bandwidth utilization with the maximum bandwidth utilization of the SLA performance evaluation data in the raw path data.
除此之外，当SLA性能评估数据中包括最大双向时延时，还可以将双向时延与该最大双向时延进行比较。In addition, when the SLA performance evaluation data includes the maximum two-way delay, the two-way delay may also be compared with the maximum two-way delay.
在一些实施例中，根据该比较方式可以得到数据比较结果。In some embodiments, the data comparison result can be obtained according to the comparison manner.
当丢包率为100％时，可以确定原始网络路径出现中断情况，因此原始网络路径异常。当丢包率不为100％，但是其他原始数据特征超过原始路径数据时，确定原始网络路径出现质差情况，因此原始网络路径异常。When the packet loss rate is 100%, it can be determined that the original network path is interrupted, so the original network path is abnormal. When the packet loss rate is not 100%, but other original data characteristics exceed the original path data, it is determined that the original network path has poor quality, and therefore the original network path is abnormal.
除此之外，原始网络路径的中断情况也可以是采集到用户投诉的中断信息确定的，本示例性实施例对此不做特殊限定。In addition, the interruption of the original network path may also be determined by collecting interruption information complained by users, which is not specifically limited in this exemplary embodiment.
值得说明的是，也可以不把原始路径数据作为原始数据特征比较的门限值，该门限值还可以是根据历史情况设定的，例如带宽利用率门限值为80％，时延10ms等，本示例性实施例对此不做特殊限定。It is worth noting that the original path data may not be used as the threshold value for feature comparison of the original data, and the threshold value may also be set according to historical conditions, for example, the bandwidth utilization threshold value is 80%, and the delay is 10ms etc., which is not specifically limited in this exemplary embodiment.
在一些实施例中，当确定原始网络路径异常时，可以获取原始路径数据中的原始流量数据和备选路径数据中的备选流量数据。原始流量数据包括原始路径数据中的上行流量和下行流量，备选流量数据包括备选路径数据中的上行流量和下行流量。In some embodiments, when it is determined that the original network path is abnormal, the original flow data in the original path data and the candidate flow data in the candidate path data may be acquired. The original traffic data includes uplink traffic and downlink traffic in the original path data, and the candidate traffic data includes uplink traffic and downlink traffic in the candidate path data.
当没有原始数据特征超过限定值时，可以确定原始网络路径正常，无需获取原始路径数据中的原始流量数据和备选路径数据中的备选流量数据进行后续处理。When no characteristic of the original data exceeds the limit value, it can be determined that the original network path is normal, and there is no need to obtain the original flow data in the original path data and the candidate flow data in the candidate path data for subsequent processing.
网络流量仿真模块是在进行备选路径评估和备选路径推荐之前，需要判断备选网络路径目前承载的报文类型和原始网络路径是否相同，然后在备选网络路径的基础上叠加原始网络路径的流量数据进行备选路径网络流量仿真。Before evaluating and recommending alternative paths, the network traffic simulation module needs to judge whether the packet type currently carried by the alternative network path is the same as that of the original network path, and then superimpose the original network path on the basis of the alternative network path The flow data of the alternative path network flow simulation is carried out.
图8示出了网络流量仿真模块的处理方法的流程示意图，如图8所示：Figure 8 shows a schematic flow chart of the processing method of the network traffic simulation module, as shown in Figure 8:
在步骤S810中，判断报文类型是否相同。In step S810, it is determined whether the message types are the same.
对原始路径数据中的原始类型数据和备选路径数据中的备选类型数据进行报文类型比较得到类型比较结果。The packet type comparison is performed on the original type data in the original route data and the alternative type data in the alternative route data to obtain a type comparison result.
获取原始路径数据中的报文类型作为原始类型数据，并获取备选路径数据中的报文类型作为备选类型数据。进一步的，可以判断原始类型数据与备选类型数据是否相同，以得到类型比较结果。The packet type in the original path data is obtained as the original type data, and the packet type in the alternate path data is obtained as the alternate type data. Further, it can be judged whether the original type data is the same as the alternative type data, so as to obtain a type comparison result.
在步骤S820中，主线路与备选线路流量叠加。In step S820, the traffic of the main line and the backup line are superimposed.
若类型比较结果为原始类型数据与备选类型数据相同，对原始流量数据和备选流量数据进行流量叠加处理得到叠加流量数据。If the type comparison result shows that the original type data is the same as the alternative type data, the original traffic data and the alternative traffic data are subjected to traffic superposition processing to obtain superimposed traffic data.
当原始类型数据与备选类型数据相同时，表明原始路径数据的上行流量和下行流量，以及备选路径数据的上行流量和下行流量是相同的，因此，可以对原始流量数据和备选流量数据进行流量叠加处理得到叠加流量数据。When the original type data is the same as the alternate type data, it indicates that the upstream traffic and downstream traffic of the original path data and the upstream traffic and downstream traffic of the alternate path data are the same, therefore, the original traffic data and the alternate traffic data can be compared Perform traffic superposition processing to obtain superimposed traffic data.
值得说明的是，流量叠加处理时，可以是原始路径数据中的上行流量与备选路径数据的上行流量进行叠加，原始路径数据的下行流量和备选路径数据中的下行流量进行叠加。因此，叠加流量数据可以包括流量叠加后的上行流量和叠加后的下行流量。It is worth noting that, during traffic superposition processing, the uplink traffic in the original path data and the uplink traffic in the alternative path data may be superimposed, and the downlink traffic in the original path data and the downlink traffic in the alternative path data may be superimposed. Therefore, the superimposed traffic data may include superimposed uplink traffic and superimposed downlink traffic.
在步骤S830中，计算备选线路报文长度。In step S830, the length of the candidate line message is calculated.
若类型比较结果为原始类型数据与备选类型数据不同，对原始路径数据和备选路径数据进行叠加更新计算得到叠加流量数据。If the type comparison result shows that the original type data is different from the alternative type data, the original path data and the alternate path data are superimposed and updated to obtain superimposed traffic data.
当原始类型数据与备选类型数据不同时，表明原始路径数据的上行流量和下行流量，以及备选路径数据的上行流量和下行流量是不同的，因此，可以对原始流量数据和备选流量数据进行流量更新计算得到叠加流量数据。When the original type data is different from the alternative type data, it indicates that the uplink traffic and downlink traffic of the original path data and the uplink traffic and downlink traffic of the alternative path data are different, therefore, the original traffic data and the alternative traffic data can be compared Perform flow update calculations to obtain superimposed flow data.
在一些实施例中，可以首先计算备选路径报文长度。该备选路径报文长度＝备选路径流量/备选路径报文数量。In some embodiments, the packet length of the alternative path may be calculated first. The packet length of the alternative path=traffic of the alternate path/number of packets of the alternate path.
在步骤S840中，计算该业务在备选线路的流量。In step S840, the traffic of the service on the alternative line is calculated.
在计算出备选路径报文长度之后，可以计算新业务流量。该新业务流量＝新业务报文数量×备选路径报文长度。After calculating the packet length of the alternative path, the new service flow can be calculated. The new service flow=number of new service packets×length of alternative path packets.
由于是将原始网络路径的上行流量和下行流量叠加至备选网络路径的上行流量和下行流量中，因此，该新业务流量即为叠加后的叠加流量数据。Since the uplink traffic and downlink traffic of the original network path are superimposed into the uplink traffic and downlink traffic of the alternative network path, the new service traffic is superimposed superimposed traffic data.
在计算备选路径报文长度时所使用的备选路径流量可以是备选网络路径的上行流量或下行流量，所使用的备选路径报文数量可以是备选网络路径的发送报文数量或接收报文数量，本示例性实施例对此不做特殊限定。The alternative path traffic used when calculating the length of the alternative path message can be the upstream or downstream traffic of the alternative network path, and the number of alternative path packets used can be the number of packets sent by the alternative network path or The number of received packets is not specifically limited in this exemplary embodiment.
在一些实施例中，完成当原始类型数据与备选类型数据不同时的流量叠加处理。In some embodiments, traffic overlay processing is done when the original type data is different from the alternative type data.
备选路径评估模块在备选网络路径更新流量后，计算备选网络路径的带宽利用率，评估收发业务流时间差等。The alternative path evaluation module calculates the bandwidth utilization rate of the alternative network path and evaluates the time difference between sending and receiving service flows after the alternative network path updates the traffic.
在得到叠加流量数据之后，可以进一步对叠加流量数据进行路径评估处理。After the superimposed traffic data is obtained, path evaluation processing may be further performed on the superimposed traffic data.
获取备选路径数据中的网络路径数据，并对叠加流量数据和网络路径数据进行路 径评估处理得到网络路径数据。Obtain the network path data in the candidate path data, and perform path evaluation processing on the superimposed traffic data and network path data to obtain the network path data.
为进行路径评估处理，可以获取到的备选路径数据中的网络路径数据可以包括备选网络路径中的路由个数、带宽、链路光纤总长度，也可以包括其他能够进行路径评估处理的数据，本示例性实施例对此不做特殊限定。For path evaluation processing, the network path data in the candidate path data that can be obtained may include the number of routes in the candidate network path, the bandwidth, the total length of link fibers, and other data that can be processed for path evaluation , which is not specifically limited in this exemplary embodiment.
在一些实施例中，可以对叠加流量数据和网络路径数据进行路径评估处理。In some embodiments, path evaluation processing may be performed on overlay traffic data and network path data.
例如，路由跳数＝路径中路由个数-1，带宽利用率＝备选路径总流量/备选路径带宽，备选路径长度＝备选路径的链路光纤总长度。该路由跳数、带宽利用率和备选路径长度即为网络评估数据。For example, the number of route hops=the number of routes in the path-1, the bandwidth utilization rate=the total traffic of the alternative path/the bandwidth of the alternative path, and the length of the alternative path=the total length of link fibers of the alternative path. The route hop count, bandwidth utilization rate and candidate path length are the network evaluation data.
图9示出了备选路径评估模块预测目标预测数据的方法的流程示意图，如图9所示：Fig. 9 shows a schematic flowchart of a method for predicting target prediction data by the alternative path evaluation module, as shown in Fig. 9:
在步骤S910中，输入备选路径历史流量、时间戳、单向时延和丢包率。In step S910, the historical traffic, time stamp, one-way delay and packet loss rate of the alternative path are input.
获取当前时间戳和备选网络路径的历史预测数据，并利用当前时间戳和历史预测数据进行时延预测处理得到目标预测数据，以确定网络评估数据和目标预测数据为路径评估数据。Obtain the current time stamp and historical forecast data of alternative network paths, and use the current time stamp and historical forecast data to perform delay prediction processing to obtain target forecast data, so as to determine the network evaluation data and target forecast data as path evaluation data.
在一些实施例中，该历史预测数据可以包括历史流量数据和丢包率。例如，历史流量数据可以是以当前时间为节点，汇总的当前时间之前的所有流量数据。In some embodiments, the historical forecast data may include historical traffic data and packet loss rate. For example, the historical traffic data may take the current time as a node and summarize all the traffic data before the current time.
在一些实施例中，可以利用当前时间戳和历史预测数据进行时延预测处理。In some embodiments, the latency prediction process can be performed using the current timestamp and historical prediction data.
例如，将当前时间戳和历史预测数据输入至预先训练好的时延预测模型中，以使时延预测模型输出目标预测数据。For example, the current time stamp and historical prediction data are input into a pre-trained delay prediction model, so that the delay prediction model outputs target prediction data.
目标预测数据包括单向时延数据和预测的丢包率。单向时延数据即为收发业务流时间差。该收发业务流时间差＝目的端接收业务报文时间-源端发送业务报文时间。The target prediction data includes one-way delay data and predicted packet loss rate. The one-way delay data is the time difference between sending and receiving service flows. The time difference between sending and receiving service flows = the time when the destination end receives the service message - the time when the source end sends the service message.
例如，该时延预测模型可以是LightGBM模型。例如，获取备选网络路径的历史流量样本、时间戳样本、单向时延样本和丢包率样本。For example, the delay prediction model may be a LightGBM model. For example, obtain historical traffic samples, time stamp samples, one-way delay samples, and packet loss rate samples of alternative network paths.
在步骤S920中，分为训练集或测试集。In step S920, it is divided into training set or test set.
将样本数据按照7：3的比例分为训练集和测试集。The sample data is divided into training set and test set according to the ratio of 7:3.
在步骤S930中，基于训练集训练LightGBM算法模型。In step S930, the LightGBM algorithm model is trained based on the training set.
在一些实施例中，基于训练集训练LightGBM模型，并基于测试集对LightGBM模型的参数进行调整。In some embodiments, the LightGBM model is trained based on the training set, and the parameters of the LightGBM model are adjusted based on the test set.
首先，输入历史流量样本、时间戳样本、单向时延样本和丢包率样本，并初始化f 0(x i)；然后，拟合残差树T(x；θ m)，并基于公式(3)和公式(4)更新公式(2)，以此 类推，不断重复拟合步骤和更新步骤进行迭代直至收敛，以得到训练好的LightGBM模型。 First, input historical traffic samples, time stamp samples, one-way delay samples and packet loss rate samples, and initialize f 0 (xi ) ; then, fit the residual tree T(x; θ m ), and based on the formula ( 3) and formula (4) to update formula (2), and so on, repeating the fitting step and updating step until convergence, so as to obtain the trained LightGBM model.
在步骤S940中，预测叠加流量后的备选路径单向时延和丢包率。In step S940, the one-way time delay and packet loss rate of the alternative path after superimposing traffic are predicted.
在训练好LightGBM模型之后，可以将当前时间戳和历史预测数据输入至已训练好的LightGBM模型中进行目标预测数据的预测，因此，该LightGBM模型可以输出预测出的单向时延数据和丢包率作为目标预测数据。After the LightGBM model is trained, the current timestamp and historical forecast data can be input into the trained LightGBM model to predict the target forecast data. Therefore, the LightGBM model can output the predicted one-way delay data and packet loss rate as the target forecast data.
因此，在得到目标预测数据之后，可以确定该目标预测数据和网络评估数据为路径评估数据。Therefore, after the target prediction data is obtained, the target prediction data and network evaluation data can be determined as path evaluation data.
备选路径推荐模块分别基于最小带宽利用率、最小路由跳数、最短时间、最短路径长度和最小丢包率，以及五者的综合指标进行备选网络路径的推荐，以便用户可以根据自身喜好及对业务的需求选择备选网络路径。The alternative path recommendation module recommends alternative network paths based on the minimum bandwidth utilization, the minimum routing hops, the shortest time, the shortest path length, the minimum packet loss rate, and the comprehensive indicators of the five, so that users can choose according to their own preferences and Select an alternative network path based on business requirements.
图10示出了备选路径推荐模块的处理方法的流程示意图，如图10所示：Fig. 10 shows a schematic flow chart of the processing method of the alternative path recommendation module, as shown in Fig. 10:
在步骤S1010中，计算路径长度、路由跳数、单向时延、带宽利用率和丢包率。In step S1010, path length, route hop count, one-way delay, bandwidth utilization rate and packet loss rate are calculated.
在步骤S1020中，与SLA最大带宽利用率、最大单向时延和最大丢包率进行比值计算。In step S1020, a ratio calculation is performed with the SLA maximum bandwidth utilization rate, the maximum one-way delay and the maximum packet loss rate.
对备选路径数据进行数据清洗处理得到备选清洗数据，并对备选清洗数据进行数据特征计算得到备选数据特征。Perform data cleaning processing on the candidate path data to obtain candidate cleaning data, and perform data characteristic calculation on the candidate cleaning data to obtain candidate data characteristics.
在一些实施例中，数据清洗处理可以包括备选路径数据的单位统一、缺失值处理和异常值处理等，也可以包括其他处理方式，本示例性实施例对此不做特殊限定。In some embodiments, the data cleaning process may include unit unification of candidate path data, missing value processing, outlier value processing, etc., and may also include other processing methods, which are not specifically limited in this exemplary embodiment.
对于数据清洗处理后的备选清洗数据，还可以进行小时的流量峰值计算，以统一按照小时粒度的接收报文数量和发送报文数量等。For the alternative cleaning data after data cleaning processing, hourly traffic peak calculations can also be performed to unify the number of received packets and the number of sent packets at an hourly granularity.
在一些实施例中，还可以对备选清洗数据进行数据特征计算得到备选数据特征。In some embodiments, data feature calculation may also be performed on the candidate cleaning data to obtain candidate data features.
在一些实施例中，该备选数据特征可以包括单向时延、双向时延、丢包率和带宽利用率。In some embodiments, the candidate data characteristics may include one-way delay, two-way delay, packet loss rate, and bandwidth utilization.
例如，单向时延＝Egress节点报文接收时间-Ingress节点报文发送时间；双向时延＝(Egress节点报文接收时间-Ingress节点报文发送时间)+Ingress节点报文接收时间-Egress节点报文发送时间)；丢包率＝(Egress节点接收报文数-Ingress节点发送报文数)*100％/Ingress节点发送报文数；带宽利用率＝流量/带宽。For example, one-way delay=Egress node message receiving time-Ingress node message sending time; two-way delay=(Egress node message receiving time-Ingress node message sending time)+Ingress node message receiving time-Egress node message sending time); packet loss rate=(the number of messages received by the Egress node-the number of messages sent by the Ingress node)*100%/the number of messages sent by the Ingress node; bandwidth utilization rate=traffic/bandwidth.
在一些实施例中，可以对备选特征数据和特征门限数据进行比较。In some embodiments, the candidate feature data can be compared to feature threshold data.
例如，从备选清洗数据中获取到的特征门限数据可以包括备选网络路径的SLA性 能评估数据，亦即最大丢包率、最大单向时延和最大带宽利用率。For example, the characteristic threshold data obtained from the candidate cleaning data may include the SLA performance evaluation data of the candidate network path, that is, the maximum packet loss rate, the maximum one-way delay and the maximum bandwidth utilization.
在一些实施例中，可以是对备选特征数据和特征门限数据进行除法计算得到二者比值，以作为备选比较结果。In some embodiments, a ratio between the candidate characteristic data and the characteristic threshold data may be calculated to obtain a ratio of the two as the candidate comparison result.
在步骤S1030中，舍弃备选路径。In step S1030, the alternative path is discarded.
当备选比较结果为备选特征数据和特征门限数据的比值大于1时，可以将该备选网络路径舍弃，对其他备选网络路径的路径评估数据和网络偏好数据进行路径推荐处理。When the candidate comparison result is that the ratio of the candidate feature data to the feature threshold data is greater than 1, the candidate network path can be discarded, and path recommendation processing is performed on the path evaluation data and network preference data of other candidate network paths.
在步骤S1040中，计算Z-score值。In step S1040, calculate the Z-score value.
当备选比较结果为备选特征数据和特征门限数据的比值小于或等于1时，可以继续对备选网络路径的路径评估数据和网络偏好数据进行路径推荐处理。When the candidate comparison result is that the ratio of the candidate feature data to the feature threshold data is less than or equal to 1, the path recommendation process may continue to be performed on the path evaluation data and network preference data of the candidate network path.
对路径评估数据进行数据标准化得到标准评估数据，并对标准评估数据和网络偏好数据进行路径推荐处理得到与备选网络路径对应的路径推荐参数。Data standardization is performed on the path evaluation data to obtain standard evaluation data, and path recommendation processing is performed on the standard evaluation data and network preference data to obtain path recommendation parameters corresponding to candidate network paths.
在一些实施例中，可以按照公式(5)利用Z-score算法对路径评估数据进行数据标准化得到标准评估数据。In some embodiments, the Z-score algorithm may be used to standardize the path evaluation data according to formula (5) to obtain standard evaluation data.
例如，当有5条备选网络路径，且计算备选路径长度时，μ为5条备选网络路径的备选路径长度均值，σ为5条备选网络路径的备选路径长度的标准差。以此类推，也可以计算另外三项的路径评估数据得到对应的标准评估数据。For example, when there are 5 candidate network paths and the length of the candidate paths is calculated, μ is the mean value of the candidate path lengths of the 5 candidate network paths, and σ is the standard deviation of the candidate path lengths of the 5 candidate network paths . By analogy, the path evaluation data of the other three items can also be calculated to obtain the corresponding standard evaluation data.
在步骤S1050中，统计用户对五项指标的喜好占比。In step S1050, statistics are made on the percentages of preferences of the users for the five indicators.
为使最优备选路径的确定可以综合考虑用户喜好，因此还可以进一步获取网络偏好数据。In order to make the determination of the optimal alternative path comprehensively consider user preferences, network preference data can also be further obtained.
该网络偏好数据是用户对路径评估数据中路由跳数、带宽利用率、备选路径长度、单向时延数据和丢包率的偏好数据。标准评估数据中备选路径长度、路由跳数、单向时延数据、带宽利用率和丢包率的网络偏好数据分别为为w 1、w 2、w 3、w 4和w 5。并且，由于五项指标越小，路径最优，因此五项网络偏好数据都小于0。 The network preference data is the user's preference data on the route hop count, bandwidth utilization rate, candidate path length, one-way delay data and packet loss rate in the path evaluation data. The network preference data of the alternative path length, routing hops, one-way delay data, bandwidth utilization rate and packet loss rate in the standard evaluation data are w 1 , w 2 , w 3 , w 4 and w 5 respectively. Moreover, since the smaller the five indexes are, the path is optimal, so the five network preference data are all less than 0.
举例而言，在n个用户中，30％偏好短路径、40％偏好少跳数、10％偏好短时延、10％偏好少带宽利用率、10％偏好较低丢包率，因此，可以得到网络偏好数据w 1＝-0.3，w 2＝-0.4，w 3＝-0.1，w 4＝-0.1，w 5＝-0.1。 For example, among n users, 30% prefer short paths, 40% prefer less hops, 10% prefer short delays, 10% prefer less bandwidth utilization, and 10% prefer lower packet loss rates. Therefore, it can The network preference data w 1 =-0.3, w 2 =-0.4, w 3 =-0.1, w 4 =-0.1, w 5 =-0.1 are obtained.
在步骤S1060中，计算备选路径推荐度y值。In step S1060, the value y of the recommendation degree of the alternative path is calculated.
在一些实施例中，对标准评估数据和网络偏好数据进行路径推荐处理时，可以按照公式(6)，以网络偏好数据作为权重进行计算得到路径推荐参数y。In some embodiments, when the path recommendation process is performed on the standard evaluation data and the network preference data, the path recommendation parameter y can be obtained by calculating with the network preference data as a weight according to formula (6).
在步骤S1070中，输出值最大的备选路径编号。In step S1070, the number of the candidate path with the largest value is output.
在得到路径推荐参数之后，可以对多条备选网络路径的路径推荐参数进行比较，以确定路径推荐参数最大的备选网络路径作为原始网络路径的最优备选路径。After the path recommendation parameters are obtained, the path recommendation parameters of multiple candidate network paths can be compared to determine the candidate network path with the largest path recommendation parameter as the optimal candidate path of the original network path.
在该应用场景下，还可以利用一示例说明网络路径的处理方法。In this application scenario, an example may also be used to illustrate the processing method of the network path.
在一些实施例中，原始网络路径的报文类型为IPV6(Internet Protocol Version 6，互联网协议第6版)，备选网络路径1中无报文，备选网络路径2中报文类型为IPV6，备选网络路径3中报文类型为SRV6。In some embodiments, the packet type of the original network path is IPV6 (Internet Protocol Version 6, Internet Protocol Version 6), there is no packet in the alternative network path 1, and the packet type in the alternate network path 2 is IPV6, The packet type in the candidate network path 3 is SRV6.
采集2021年原始网络路径和备选网络路径经过各网络路由器端口的发送报文数量、接收报文数量、上行流量、下行流量、带宽(2M)和业务的SLA性能评估数据，该SLA性能评估数据包括最大丢包率、最大单向时延和最大带宽利用率。并且，还可以获取用户的业务路径喜好数据，亦即网络偏好数据。Collect the SLA performance evaluation data of the number of sent packets, received packets, upstream traffic, downstream traffic, bandwidth (2M) and services of the original network path and the alternative network path through each network router port in 2021. The SLA performance evaluation data Including maximum packet loss rate, maximum one-way delay and maximum bandwidth utilization. In addition, the service path preference data of the user, that is, network preference data, may also be obtained.
对于数据清洗处理后的性能数据，还可以进行小时的流量峰值计算，以统一按照小时粒度的接收报文数量和发送报文数量等。For the performance data after data cleaning and processing, hourly traffic peak calculations can also be performed to unify the number of received packets and the number of sent packets at an hourly granularity.
进一步的，还可以对原始清洗数据进行数据特征计算得到原始数据特征。该原始特征数据可以包括单向时延、双向时延、丢包率和带宽利用率。Further, data feature calculation can also be performed on the original cleaned data to obtain the original data features. The original characteristic data may include one-way delay, two-way delay, packet loss rate and bandwidth utilization rate.
基于性能数据(丢包率、单向时延、双向时延和带宽利用率)以及SLA数据(最大丢包率、最大单向时延、最大双向时延和最大带宽利用率)进行端到端业务质量评估。End-to-end based on performance data (packet loss rate, one-way delay, two-way delay, and bandwidth utilization) and SLA data (maximum packet loss, maximum one-way delay, maximum two-way delay, and maximum bandwidth utilization) Business quality assessment.
业务质量评估的评估结果如表(1)所示：The evaluation results of service quality evaluation are shown in Table (1):
Figure PCTCN2022112000-appb-000005
Figure PCTCN2022112000-appb-000005
表(1)Table 1)
在一些实施例中，判断原始网络路径和备选网络路径的报文类型是否相同。In some embodiments, it is determined whether the packet types of the original network path and the candidate network path are the same.
若类型比较结果为原始类型数据与备选类型数据相同，对原始流量数据和备选流量数据进行流量叠加处理得到叠加流量数据。If the type comparison result shows that the original type data is the same as the alternative type data, the original traffic data and the alternative traffic data are subjected to traffic superposition processing to obtain superimposed traffic data.
若类型比较结果为原始类型数据与备选类型数据不同，对原始路径数据和备选路径数据进行叠加更新计算得到叠加流量数据。If the type comparison result shows that the original type data is different from the alternative type data, the original path data and the alternate path data are superimposed and updated to obtain superimposed traffic data.
计算报文长度，进而分别计算该业务加入到备选网络路径1、备选网络路径2和备 选网络路径3的流量中的结果如表(2)所示：Calculate the length of the message, and then calculate the results of adding the service to the traffic of alternative network path 1, alternative network path 2, and alternative network path 3, as shown in Table (2):
Figure PCTCN2022112000-appb-000006
Figure PCTCN2022112000-appb-000006
表(2)Table 2)
统计备选网络路径1、备选网络路径2和备选网络路径3的备选路径长度、路由跳数，并基于叠加后的流量计算带宽利用率、预测单向时延数据和丢包率。Count the candidate path lengths and routing hops of candidate network path 1, candidate network path 2, and candidate network path 3, and calculate bandwidth utilization, predict one-way delay data, and packet loss rate based on superimposed traffic.
进而，获取叠加后的SLA最大带宽利用率、最大单向时延和最大丢包率。Furthermore, the maximum bandwidth utilization rate, the maximum one-way delay and the maximum packet loss rate of the superimposed SLA are obtained.
将叠加后的流量计算带宽利用率、预测的单向时延和丢包率与叠加后的SLA最大带宽利用率、最大单向时延和最大丢包率进行对比，备选网络路径1、备选网络路径2和备选网络路径3并无超限情况。Compare the superimposed traffic calculation bandwidth utilization, predicted one-way delay, and packet loss rate with the superimposed SLA maximum bandwidth utilization, maximum one-way delay, and maximum packet loss rate. The selected network path 2 and the alternative network path 3 are not overrun.
最后，计算备选网络路径1、备选网络路径2和备选网络路径3的备选路径长度、路由跳数、基于叠加后的流量计算带宽利用率、预测的单向时延和预测的丢包率的Z-score值，并确定五项指标权重值，亦即网络偏好数据，以根据网络偏好数据计算y值。Finally, calculate the candidate path lengths, route hops, bandwidth utilization rate, predicted one-way delay and predicted loss based on the superimposed traffic The Z-score value of the packet rate, and determine the weight value of the five indicators, that is, the network preference data, so as to calculate the y value based on the network preference data.
根据y值得出最终备选网络路径的推荐排序如表(3)所示：According to the value of y, the recommended ranking of the final candidate network paths is shown in Table (3):
Figure PCTCN2022112000-appb-000007
Figure PCTCN2022112000-appb-000007
表(3)table 3)
因此，最终选择备选网络路径3作为原始网络路径的最优备选路径。Therefore, the candidate network path 3 is finally selected as the optimal candidate path of the original network path.
在该应用场景下的网络路径的处理方法，一方面，对原始流量数据和备选流量数据进行流量叠加处理，考虑了流量叠加过程中原始流量数据与备选流量数据不同的情况，能够更加精确地进行最优备选路径的选取，提升了确定的最优备选路径的准确度；另一方面，利用网络偏好数据和路径评估数据进行路径推荐处理，综合考虑多项影响路径选取的性能指标，从多个角度提升网络运维效率，同时满足用户在不同情况下的路径选择需求，适应性更强，灵活性更佳。The network path processing method in this application scenario, on the one hand, performs traffic superposition processing on the original traffic data and the alternative traffic data, and considers the difference between the original traffic data and the alternative traffic data in the traffic superposition process, which can be more accurate The selection of the optimal alternative path improves the accuracy of the determined optimal alternative path; on the other hand, the network preference data and path evaluation data are used for path recommendation processing, and multiple performance indicators that affect path selection are comprehensively considered. , improve the efficiency of network operation and maintenance from multiple perspectives, and at the same time meet the path selection needs of users in different situations, with stronger adaptability and better flexibility.
此外，在本公开的示例性实施例中，还提供一种网络路径的处理装置。图11示出了网络路径的处理装置的结构示意图，如图11所示，网络路径的处理装置1100可以包括：流量叠加模块1110、路径评估模块1120和路径推荐模块1130。其中：In addition, in an exemplary embodiment of the present disclosure, a device for processing a network path is also provided. FIG. 11 shows a schematic structural diagram of a network path processing device. As shown in FIG. 11 , the network path processing device 1100 may include: a traffic superposition module 1110 , a path evaluation module 1120 and a path recommendation module 1130 . in:
流量叠加模块1110，被配置为获取原始网络路径的原始流量数据和备选网络路径的备选流量数据，对原始流量数据和备选流量数据进行流量叠加处理得到叠加流量数据；路径评估模块1120，被配置为获取备选网络路径的网络路径数据，并对叠加流量数据和网络路径数据进行路径评估处理得到路径评估数据；路径推荐模块1130，被配置为获取与路径评估数据对应的网络偏好数据，对路径评估数据和网络偏好数据进行路径推荐处理，以在备选网络路径中确定原始网络路径的最优备选路径。The traffic superposition module 1110 is configured to obtain the original traffic data of the original network path and the candidate traffic data of the candidate network path, and perform traffic superposition processing on the original traffic data and the candidate traffic data to obtain the superimposed traffic data; the path evaluation module 1120, configured to acquire network path data of alternative network paths, and perform path evaluation processing on the superimposed traffic data and network path data to obtain path evaluation data; the path recommendation module 1130 is configured to obtain network preference data corresponding to the path evaluation data, Path recommendation processing is performed on the path evaluation data and network preference data to determine the optimal alternative path of the original network path among the alternative network paths.
在本发明的一种示例性实施例中，所述管理配置信息，包括：所述获取原始网络路径的原始流量数据和备选网络路径的备选流量数据，包括：In an exemplary embodiment of the present invention, the management configuration information includes: the acquisition of the original traffic data of the original network path and the alternative traffic data of the alternative network path includes:
获取原始网络路径的原始路径数据和备选网络路径的备选路径数据，并对所述原始路径数据进行数据清洗处理得到原始清洗数据；Obtaining the original path data of the original network path and the candidate path data of the candidate network path, and performing data cleaning processing on the original path data to obtain the original cleaning data;
对所述原始清洗数据进行数据特征计算得到原始数据特征，并对所述原始数据特征和所述原始路径数据进行比较得到数据比较结果；performing data feature calculation on the original cleaning data to obtain original data features, and comparing the original data features with the original path data to obtain a data comparison result;
若所述数据比较结果为所述原始数据特征未满足所述原始路径数据中包括的服务级别协议SLA性能评估数据的要求，确定所述原始网络路径异常，并获取所述原始路径数据中的原始流量数据和备选路径数据中的备选流量数据。If the result of the data comparison is that the characteristics of the original data do not meet the requirements of the service level agreement (SLA) performance evaluation data included in the original path data, determine that the original network path is abnormal, and obtain the original path data in the original path. Alternate flow data in flow data and alternative path data.
在本发明的一种示例性实施例中，所述对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据，包括：In an exemplary embodiment of the present invention, performing traffic superposition processing on the original traffic data and the candidate traffic data to obtain superimposed traffic data includes:
对所述原始路径数据中的原始类型数据和所述备选路径数据中的备选类型数据进行报文类型比较得到类型比较结果；performing message type comparison on the original type data in the original route data and the alternative type data in the alternative route data to obtain a type comparison result;
若所述类型比较结果为所述原始类型数据与所述备选类型数据相同，对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据；If the type comparison result is that the original type data is the same as the alternative type data, performing traffic superposition processing on the original traffic data and the alternative traffic data to obtain superimposed traffic data;
若所述类型比较结果为所述原始类型数据与所述备选类型数据不同，对所述原始路径数据和所述备选路径数据进行叠加更新计算得到叠加流量数据。If the type comparison result is that the original type data is different from the alternative type data, superimposing and updating the original path data and the alternative path data is performed to obtain superimposed traffic data.
在本发明的一种示例性实施例中，所述获取所述备选网络路径的网络路径数据，并对所述叠加流量数据和所述网络路径数据进行路径评估处理得到路径评估数据，包括：In an exemplary embodiment of the present invention, the acquiring the network path data of the candidate network path, and performing path evaluation processing on the superimposed traffic data and the network path data to obtain path evaluation data includes:
获取所述备选路径数据中的网络路径数据，并对所述叠加流量数据和所述网络路 径数据进行路径评估处理得到网络路径数据；Obtaining network path data in the candidate path data, and performing path evaluation processing on the superimposed traffic data and the network path data to obtain network path data;
获取当前时间戳和所述备选网络路径的历史预测数据，并利用所述当前时间戳和所述历史预测数据进行时延预测处理得到目标预测数据，以确定所述网络路径数据和所述目标预测数据为路径评估数据。Acquiring the current time stamp and historical forecast data of the candidate network path, and using the current time stamp and the historical forecast data to perform delay prediction processing to obtain target forecast data, so as to determine the network path data and the target Prediction data is path evaluation data.
在本发明的一种示例性实施例中，所述利用所述当前时间戳和所述历史流量数据进行时延预测处理得到单向时延数据，包括：In an exemplary embodiment of the present invention, the one-way delay data obtained by using the current time stamp and the historical traffic data to perform delay prediction processing includes:
将所述当前时间戳和所述历史预测数据输入至预先训练好的时延预测模型中，以使所述时延预测模型输出目标预测数据。The current time stamp and the historical prediction data are input into a pre-trained delay prediction model, so that the delay prediction model outputs target prediction data.
在本发明的一种示例性实施例中，所述对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径，包括：In an exemplary embodiment of the present invention, the path recommendation process is performed on the path evaluation data and the network preference data, so as to determine the optimal alternative of the original network path among the candidate network paths. Choose a path, including:
对所述备选路径数据进行数据清洗处理得到备选清洗数据，并对所述备选清洗数据进行数据特征计算得到备选数据特征；performing data cleaning processing on the candidate path data to obtain candidate cleaning data, and performing data characteristic calculation on the candidate cleaning data to obtain candidate data characteristics;
获取所述备选清洗数据中的特征门限数据，并对所述备选特征数据和所述特征门限数据进行比较到备选比较结果；Obtain feature threshold data in the candidate cleaning data, and compare the candidate feature data with the feature threshold data to obtain a candidate comparison result;
基于所述备选比较结果，对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径。Based on the candidate comparison result, perform route recommendation processing on the route evaluation data and the network preference data, so as to determine the optimal candidate route of the original network route among the candidate network routes.
在本发明的一种示例性实施例中，所述对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径，包括：In an exemplary embodiment of the present invention, the path recommendation process is performed on the path evaluation data and the network preference data, so as to determine the optimal alternative of the original network path among the candidate network paths. Choose a path, including:
对所述路径评估数据进行数据标准化得到标准评估数据，并对所述标准评估数据和所述网络偏好数据进行路径推荐处理得到与所述备选网络路径对应的路径推荐参数；performing data standardization on the path evaluation data to obtain standard evaluation data, and performing path recommendation processing on the standard evaluation data and the network preference data to obtain path recommendation parameters corresponding to the candidate network paths;
对所述路径推荐参数进行推荐程度识别，以在所述备选网络路径中确定所述原始网络路径的最优备选路径。The recommendation degree identification is carried out on the path recommendation parameters, so as to determine the optimal candidate path of the original network path among the candidate network paths.
上述网络路径的处理装置1100的具体细节已经在对应的网络路径的处理方法中进行了详细的描述，因此此处不再赘述。The specific details of the foregoing network path processing apparatus 1100 have been described in detail in the corresponding network path processing method, so details are not repeated here.
应当注意，尽管在上文详细描述中提及了网络路径的处理装置1100的若干模块或者单元，但是这种划分并非强制性的。实际上，根据本公开的实施方式，上文描述的两个或更多模块或者单元的特征和功能可以在一个模块或者单元中具体化。反之，上文描述的一个模块或者单元的特征和功能可以进一步划分为由多个模块或者单元来具 体化。It should be noted that although several modules or units of the network path processing apparatus 1100 are mentioned in the above detailed description, this division is not mandatory. Actually, according to the embodiment of the present disclosure, the features and functions of two or more modules or units described above may be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided to be embodied by a plurality of modules or units.
此外，在本公开的示例性实施例中，还提供了一种能够实现上述方法的电子设备。In addition, in an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.
下面参照图12来描述根据本发明的这种实施例的电子设备1200。图12显示的电子设备1200仅仅是一个示例，不应对本发明实施例的功能和使用范围带来任何限制。An electronic device 1200 according to such an embodiment of the present invention is described below with reference to FIG. 12 . The electronic device 1200 shown in FIG. 12 is only an example, and should not limit the functions and scope of use of this embodiment of the present invention.
如图12所示，电子设备1200以通用计算设备的形式表现。电子设备1200的组件可以包括但不限于：上述至少一个处理单元1210、上述至少一个存储单元1220、连接不同***组件(包括存储单元1220和处理单元1210)的总线1230、显示单元1240。As shown in FIG. 12, electronic device 1200 takes the form of a general-purpose computing device. The components of the electronic device 1200 may include, but are not limited to: at least one processing unit 1210, at least one storage unit 1220, a bus 1230 connecting different system components (including the storage unit 1220 and the processing unit 1210), and a display unit 1240.
其中，所述存储单元存储有程序代码，所述程序代码可以被所述处理单元1210执行，使得所述处理单元1210执行本说明书上述“示例性方法”部分中描述的根据本发明各种示例性实施例的步骤。Wherein, the storage unit stores program codes, and the program codes can be executed by the processing unit 1210, so that the processing unit 1210 executes various exemplary methods according to the present invention described in the "Exemplary Methods" section of this specification. Example steps.
存储单元1220可以包括易失性存储单元形式的可读介质，例如随机存取存储单元(RAM)1221和/或高速缓存存储单元1222，还可以进一步包括只读存储单元(ROM)1223。The storage unit 1220 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 1221 and/or a cache storage unit 1222 , and may further include a read-only storage unit (ROM) 1223 .
存储单元1220还可以包括具有一组(至少一个)程序模块1225的程序/实用工具1224，这样的程序模块1225包括但不限于：操作***、一个或者多个应用程序、其它程序模块以及程序数据，这些示例中的每一个或某种组合中可能包括网络环境的实现。 Storage unit 1220 may also include a program/utility 1224 having a set (at least one) of program modules 1225, such program modules 1225 including but not limited to: an operating system, one or more application programs, other program modules, and program data, Implementations of networked environments may be included in each or some combination of these examples.
总线1230可以为表示几类总线结构中的一种或多种，包括存储单元总线或者存储单元控制器、***总线、图形加速端口、处理单元或者使用多种总线结构中的任意总线结构的局域总线。 Bus 1230 may represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local area using any of a variety of bus structures. bus.
电子设备1200也可以与一个或多个外部设备1400(例如键盘、指向设备、蓝牙设备等)通信，还可与一个或者多个使得用户能与该电子设备1200交互的设备通信，和/或与使得该电子设备1200能与一个或多个其它计算设备进行通信的任何设备(例如路由器、调制解调器等等)通信。这种通信可以通过输入/输出(I/O)接口1250进行。并且，电子设备1200还可以通过网络适配器1260与一个或者多个网络(例如局域网(LAN)，广域网(WAN)和/或公共网络，例如因特网)通信。如图所示，网络适配器1240通过总线1230与电子设备1200的其它模块通信。应当明白，尽管图中未示出，可以结合电子设备1200使用其它硬件和/或软件模块，包括但不限于：微代码、设备驱动器、冗余处理单元、外部磁盘驱动阵列、RAID***、磁带驱动器以及数据备份存储***等。The electronic device 1200 can also communicate with one or more external devices 1400 (such as keyboards, pointing devices, Bluetooth devices, etc.), and can also communicate with one or more devices that enable the user to interact with the electronic device 1200, and/or communicate with Any device (eg, router, modem, etc.) that enables the electronic device 1200 to communicate with one or more other computing devices. Such communication may occur through input/output (I/O) interface 1250 . Moreover, the electronic device 1200 can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN) and/or a public network such as the Internet) through the network adapter 1260 . As shown, the network adapter 1240 communicates with other modules of the electronic device 1200 through the bus 1230 . It should be appreciated that although not shown, other hardware and/or software modules may be used in conjunction with electronic device 1200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives And data backup storage system, etc.
通过以上的实施例的描述，本领域的技术人员易于理解，这里描述的示例实施例 可以通过软件实现，也可以通过软件结合必要的硬件的方式来实现。因此，根据本公开实施例的技术方案可以以软件产品的形式体现出来，该软件产品可以存储在一个非易失性存储介质(可以是CD-ROM，U盘，移动硬盘等)中或网络上，包括若干指令以使得一台计算设备(可以是个人计算机、服务器、终端装置、或者网络设备等)执行根据本公开实施例的方法。Through the description of the above embodiments, those skilled in the art can easily understand that the exemplary embodiments described here can be implemented by software, or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of software products, and the software products can be stored in a non-volatile storage medium (which can be CD-ROM, U disk, mobile hard disk, etc.) or on the network , including several instructions to make a computing device (which may be a personal computer, a server, a terminal device, or a network device, etc.) execute the method according to the embodiment of the present disclosure.
在本公开的示例性实施例中，还提供了一种计算机可读存储介质，其上存储有能够实现本说明书上述方法的程序产品。在一些可能的实施例中，本发明的各个方面还可以实现为一种程序产品的形式，其包括程序代码，当所述程序产品在终端设备上运行时，所述程序代码用于使所述终端设备执行本说明书上述“示例性方法”部分中描述的根据本发明各种示例性实施例的步骤。In an exemplary embodiment of the present disclosure, there is also provided a computer-readable storage medium on which a program product capable of implementing the above-mentioned method in this specification is stored. In some possible embodiments, various aspects of the present invention can also be implemented in the form of a program product, which includes program code, and when the program product is run on a terminal device, the program code is used to make the The terminal device executes the steps according to various exemplary embodiments of the present invention described in the "Exemplary Method" section above in this specification.
参考图13所示，描述了根据本发明的实施例的用于实现上述方法的程序产品1300，其可以采用便携式紧凑盘只读存储器(CD-ROM)并包括程序代码，并可以在终端设备，例如个人电脑上运行。然而，本发明的程序产品不限于此，在本文件中，可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行***、装置或者器件使用或者与其结合使用。As shown in FIG. 13 , a program product 1300 for realizing the above method according to an embodiment of the present invention is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program codes, and can be used in terminal equipment, For example running on a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium may be any tangible medium containing or storing a program, and the program may be used by or in combination with an instruction execution system, apparatus or device.
所述程序产品可以采用一个或多个可读介质的任意组合。可读介质可以是可读信号介质或者可读存储介质。可读存储介质例如可以为但不限于电、磁、光、电磁、红外线、或半导体的***、装置或器件，或者任意以上的组合。可读存储介质的更具体的例子(非穷举的列表)包括：具有一个或多个导线的电连接、便携式盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、光纤、便携式紧凑盘只读存储器(CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。The program product may reside on any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: electrical connection with one or more conductors, portable disk, hard disk, random access memory (RAM), read only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了可读程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。可读信号介质还可以是可读存储介质以外的任何可读介质，该可读介质可以发送、传播或者传输用于由指令执行***、装置或者器件使用或者与其结合使用的程序。A computer readable signal medium may include a data signal carrying readable program code in baseband or as part of a carrier wave. Such propagated data signals may take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing. A readable signal medium may also be any readable medium other than a readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.
可读介质上包含的程序代码可以用任何适当的介质传输，包括但不限于无线、有线、光缆、RF等等，或者上述的任意合适的组合。Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
可以以一种或多种程序设计语言的任意组合来编写用于执行本发明操作的程序代 码，所述程序设计语言包括面向对象的程序设计语言—诸如Java、C++等，还包括常规的过程式程序设计语言—诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算设备上执行、部分地在用户设备上执行、作为一个独立的软件包执行、部分在用户计算设备上部分在远程计算设备上执行、或者完全在远程计算设备或服务器上执行。在涉及远程计算设备的情形中，远程计算设备可以通过任意种类的网络，包括局域网(LAN)或广域网(WAN)，连接到用户计算设备，或者，可以连接到外部计算设备(例如利用因特网服务提供商来通过因特网连接)。Program code for carrying out the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages—such as Java, C++, etc., as well as conventional procedural programming languages. Programming language - such as "C" or a similar programming language. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server to execute. In cases involving a remote computing device, the remote computing device may be connected to the user computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., using an Internet service provider). business to connect via the Internet).
本领域技术人员在考虑说明书及实践这里公开的发明后，将容易想到本公开的其他实施例。本公开旨在涵盖本公开的任何变型、用途或者适应性变化，这些变型、用途或者适应性变化遵循本公开的一般性原理并包括本公开未公开的本技术领域中的公知常识或惯用技术手段。说明书和实施例仅被视为示例性的，本公开的真正范围和精神由权利要求指出。Other embodiments of the disclosure will be readily apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. The present disclosure is intended to cover any modification, use or adaptation of the present disclosure. These modifications, uses or adaptations follow the general principles of the present disclosure and include common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. . The specification and examples are to be considered exemplary only, with the true scope and spirit of the disclosure indicated by the appended claims.

Claims (15)

  1. 一种网络路径的处理方法，包括：A method for processing a network path, comprising:
    获取原始网络路径的原始流量数据和备选网络路径的备选流量数据，对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据；Acquiring original traffic data of the original network path and candidate traffic data of the candidate network path, and performing traffic superposition processing on the original traffic data and the candidate traffic data to obtain superimposed traffic data;
    获取所述备选网络路径的网络路径数据，并对所述叠加流量数据和所述网络路径数据进行路径评估处理得到路径评估数据；Acquiring network path data of the candidate network path, and performing path evaluation processing on the superimposed traffic data and the network path data to obtain path evaluation data;
    获取与所述路径评估数据对应的网络偏好数据，对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径。Acquiring network preference data corresponding to the path evaluation data, and performing path recommendation processing on the path evaluation data and the network preference data, so as to determine the optimal alternative of the original network path among the candidate network paths path.
  2. 根据权利要求1所述的网络路径的处理方法，其中，所述获取原始网络路径的原始流量数据和备选网络路径的备选流量数据包括：The method for processing a network path according to claim 1, wherein said obtaining the original traffic data of the original network path and the alternative traffic data of the alternative network path comprises:
    获取原始网络路径的原始路径数据和备选网络路径的备选路径数据，并对所述原始路径数据进行数据清洗处理得到原始清洗数据；Obtaining the original path data of the original network path and the candidate path data of the candidate network path, and performing data cleaning processing on the original path data to obtain the original cleaning data;
    对所述原始清洗数据进行数据特征计算得到原始数据特征，并对所述原始数据特征和所述原始路径数据进行比较得到数据比较结果；performing data feature calculation on the original cleaning data to obtain original data features, and comparing the original data features with the original path data to obtain a data comparison result;
    若所述数据比较结果为所述原始数据特征未满足所述原始路径数据中包括的服务级别协议SLA性能评估数据的要求，确定所述原始网络路径异常，并获取所述原始路径数据中的原始流量数据和备选路径数据中的备选流量数据。If the result of the data comparison is that the characteristics of the original data do not meet the requirements of the service level agreement (SLA) performance evaluation data included in the original path data, determine that the original network path is abnormal, and obtain the original path data in the original path. Alternate flow data in flow data and alternative path data.
  3. 根据权利要求2所述的网络路径的处理方法，其中，The processing method of the network path according to claim 2, wherein,
    所述原始数据特征包括单向时延、双向时延、丢包率和带宽利用率中的至少一项。The raw data characteristics include at least one of one-way delay, two-way delay, packet loss rate and bandwidth utilization rate.
  4. 根据权利要求3所述的网络路径的处理方法，其中，The processing method of the network path according to claim 3, wherein,
    在所述原始数据特征包括所述单向时延的情况下，所述SLA性能评估数据包括最大单向时延；In the case where the original data feature includes the one-way delay, the SLA performance evaluation data includes a maximum one-way delay;
    在所述原始数据特征包括所述双向时延的情况下，所述SLA性能评估数据包括最大双向时延；In the case where the original data feature includes the two-way delay, the SLA performance evaluation data includes a maximum two-way delay;
    在所述原始数据特征包括所述丢包率的情况下，所述SLA性能评估数据包括最大 丢包率；In the case where the original data feature includes the packet loss rate, the SLA performance evaluation data includes a maximum packet loss rate;
    在所述原始数据特征包括所述带宽利用率的情况下，所述SLA性能评估数据包括最大带宽利用率。In a case where the raw data feature includes the bandwidth utilization rate, the SLA performance evaluation data includes a maximum bandwidth utilization rate.
  5. 根据权利要求2所述的网络路径的处理方法，其中，所述对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据包括：The method for processing a network path according to claim 2, wherein performing traffic superposition processing on the original traffic data and the candidate traffic data to obtain the superimposed traffic data comprises:
    对所述原始路径数据中的原始类型数据和所述备选路径数据中的备选类型数据进行报文类型比较得到类型比较结果；performing message type comparison on the original type data in the original route data and the alternative type data in the alternative route data to obtain a type comparison result;
    若所述类型比较结果为所述原始类型数据与所述备选类型数据相同，对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据；If the type comparison result is that the original type data is the same as the alternative type data, performing traffic superposition processing on the original traffic data and the alternative traffic data to obtain superimposed traffic data;
    若所述类型比较结果为所述原始类型数据与所述备选类型数据不同，对所述原始路径数据和所述备选路径数据进行叠加更新计算得到叠加流量数据。If the type comparison result is that the original type data is different from the alternative type data, superimposing and updating the original path data and the alternative path data is performed to obtain superimposed traffic data.
  6. 根据权利要求5所述的网络路径的处理方法，其中，对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据包括：The method for processing a network path according to claim 5, wherein performing traffic superposition processing on the original traffic data and the candidate traffic data to obtain superimposed traffic data includes:
    对所述原始流量数据的上行流量和所述备选流量数据的上行流量进行流量叠加处理，得到流量叠加后的上行流量；performing traffic superposition processing on the upstream traffic of the original traffic data and the upstream traffic of the alternative traffic data, to obtain the upstream traffic after traffic superposition;
    对所述原始流量数据的下行流量和所述备选流量数据的下行流量进行流量叠加处理，得到流量叠加后的下行流量。Perform traffic superimposition processing on the downlink traffic of the original traffic data and the downlink traffic of the alternative traffic data to obtain the downlink traffic after traffic superimposition.
  7. 根据权利要求2所述的网络路径的处理方法，其中，所述获取所述备选网络路径的网络路径数据，并对所述叠加流量数据和所述网络路径数据进行路径评估处理得到路径评估数据包括：The method for processing a network path according to claim 2, wherein said acquiring the network path data of said candidate network path, and performing path evaluation processing on said superimposed flow data and said network path data to obtain path evaluation data include:
    获取所述备选路径数据中的网络路径数据，并对所述叠加流量数据和所述网络路径数据进行路径评估处理得到网络评估数据；Acquiring network path data in the candidate path data, and performing path evaluation processing on the superimposed traffic data and the network path data to obtain network evaluation data;
    获取当前时间戳和所述备选网络路径的历史预测数据，并利用所述当前时间戳和所述历史预测数据进行时延预测处理得到目标预测数据，以确定所述网络评估数据和所述目标预测数据为所述路径评估数据。Acquiring the current time stamp and historical forecast data of the candidate network path, and using the current time stamp and the historical forecast data to perform delay prediction processing to obtain target forecast data, so as to determine the network evaluation data and the target Prediction data is said path evaluation data.
  8. 根据权利要求7所述的网络路径的处理方法，其中，所述利用所述当前时间戳 和所述历史预测数据进行时延预测处理得到目标预测数据包括：The processing method of the network path according to claim 7, wherein, said using said current time stamp and said historical prediction data to perform delay prediction processing to obtain target prediction data comprises:
    将所述当前时间戳和所述历史预测数据输入至预先训练好的时延预测模型中，以使所述时延预测模型输出目标预测数据。The current time stamp and the historical prediction data are input into a pre-trained delay prediction model, so that the delay prediction model outputs target prediction data.
  9. 根据权利要求2中所述的网络路径的处理方法，其中，所述对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径包括：The method for processing a network path according to claim 2, wherein said path recommendation processing is performed on said path evaluation data and said network preference data, so as to determine said original network path among said candidate network paths The best alternative paths for include:
    对所述备选路径数据进行数据清洗处理得到备选清洗数据，并对所述备选清洗数据进行数据特征计算得到备选数据特征；performing data cleaning processing on the candidate path data to obtain candidate cleaning data, and performing data characteristic calculation on the candidate cleaning data to obtain candidate data characteristics;
    获取所述备选清洗数据中的特征门限数据，并对所述备选特征数据和所述特征门限数据进行比较到备选比较结果；Obtain feature threshold data in the candidate cleaning data, and compare the candidate feature data with the feature threshold data to obtain a candidate comparison result;
    基于所述备选比较结果，对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径。Based on the candidate comparison result, perform route recommendation processing on the route evaluation data and the network preference data, so as to determine the optimal candidate route of the original network route among the candidate network routes.
  10. 根据权利要求9所述的网络路径的处理方法，其中，The processing method of the network path according to claim 9, wherein,
    所述备选数据特征包括单向时延、双向时延、丢包率和带宽利用率中的至少一项。The candidate data characteristics include at least one of one-way delay, two-way delay, packet loss rate and bandwidth utilization rate.
  11. 根据权利要求9所述的网络路径的处理方法，其中，所述对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径包括：The method for processing a network path according to claim 9, wherein the path recommendation process is performed on the path evaluation data and the network preference data, so as to determine the original network path among the candidate network paths The best alternative paths include:
    对所述路径评估数据进行数据标准化得到标准评估数据，并对所述标准评估数据和所述网络偏好数据进行路径推荐处理得到与所述备选网络路径对应的路径推荐参数；performing data standardization on the path evaluation data to obtain standard evaluation data, and performing path recommendation processing on the standard evaluation data and the network preference data to obtain path recommendation parameters corresponding to the candidate network paths;
    对所述路径推荐参数进行推荐程度识别，以在所述备选网络路径中确定所述原始网络路径的最优备选路径。The recommendation degree identification is carried out on the path recommendation parameters, so as to determine the optimal candidate path of the original network path among the candidate network paths.
  12. 根据权利要求11所述的网络路径的处理方法，其中，The processing method of the network path according to claim 11, wherein,
    所述推荐程度与所述路径推荐参数的数值呈正相关关系。The recommendation degree is positively correlated with the value of the path recommendation parameter.
  13. 一种网络路径的处理装置，包括：A device for processing a network path, comprising:
    流量叠加模块，被配置为获取原始网络路径的原始流量数据和备选网络路径的备 选流量数据，对所述原始流量数据和所述备选流量数据进行流量叠加处理得到叠加流量数据；The traffic superposition module is configured to obtain the original traffic data of the original network path and the alternative traffic data of the alternative network path, and perform traffic superposition processing on the original traffic data and the alternative traffic data to obtain superimposed traffic data;
    路径评估模块，被配置为获取所述备选网络路径的网络路径数据，并对所述叠加流量数据和所述网络路径数据进行路径评估处理得到路径评估数据；The path evaluation module is configured to obtain network path data of the candidate network path, and perform path evaluation processing on the superimposed traffic data and the network path data to obtain path evaluation data;
    路径推荐模块，被配置为获取与所述路径评估数据对应的网络偏好数据，对所述路径评估数据和所述网络偏好数据进行路径推荐处理，以在所述备选网络路径中确定所述原始网络路径的最优备选路径。a path recommendation module, configured to acquire network preference data corresponding to the path evaluation data, and perform path recommendation processing on the path evaluation data and the network preference data, so as to determine the original network path among the candidate network paths The best alternative for network paths.
  14. 一种计算机可读存储介质，所述计算机可读存储介质上存储有计算机程序，其中，所述计算机程序被处理器执行时实现权利要求1-12中任意一项所述的网络路径的处理方法。A computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, wherein, when the computer program is executed by a processor, the method for processing a network path according to any one of claims 1-12 is implemented .
  15. 一种电子设备，包括：An electronic device comprising:
    处理器；processor;
    存储器，用于存储所述处理器的可执行指令；a memory for storing executable instructions of the processor;
    其中，所述处理器被配置为经由执行所述可执行指令来执行权利要求1-12中任意一项所述的网络路径的处理方法。Wherein, the processor is configured to execute the network path processing method according to any one of claims 1-12 by executing the executable instructions.
PCT/CN2022/112000 2021-08-12 2022-08-12 Network path processing method and apparatus, storage medium, and electronic device WO2023016539A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110924959.4A CN115914086A (en) 2021-08-12 2021-08-12 Network path processing method and device, storage medium and electronic equipment
CN202110924959.4 2021-08-12

Publications (1)

Publication Number Publication Date
WO2023016539A1 true WO2023016539A1 (en) 2023-02-16

Family

ID=85199886

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/112000 WO2023016539A1 (en) 2021-08-12 2022-08-12 Network path processing method and apparatus, storage medium, and electronic device

Country Status (2)

Country Link
CN (1) CN115914086A (en)
WO (1) WO2023016539A1 (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150319076A1 (en) * 2014-05-02 2015-11-05 Cisco Technology, Inc. Centralized predictive routing using delay predictability measurements
CN107959633A (en) * 2017-11-18 2018-04-24 浙江工商大学 A kind of load balance method based on price mechanism in industry real-time network
CN110290065A (en) * 2019-06-17 2019-09-27 南京邮电大学 SDN network routing resource based on load balancing Yu QoE metric value
CN112532518A (en) * 2020-11-26 2021-03-19 新华三技术有限公司 Path selection method and device of segment routing strategy
CN113037627A (en) * 2021-03-03 2021-06-25 烽火通信科技股份有限公司 Method and device for selecting network service line resources

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150319076A1 (en) * 2014-05-02 2015-11-05 Cisco Technology, Inc. Centralized predictive routing using delay predictability measurements
CN107959633A (en) * 2017-11-18 2018-04-24 浙江工商大学 A kind of load balance method based on price mechanism in industry real-time network
CN110290065A (en) * 2019-06-17 2019-09-27 南京邮电大学 SDN network routing resource based on load balancing Yu QoE metric value
CN112532518A (en) * 2020-11-26 2021-03-19 新华三技术有限公司 Path selection method and device of segment routing strategy
CN113037627A (en) * 2021-03-03 2021-06-25 烽火通信科技股份有限公司 Method and device for selecting network service line resources

Also Published As

Publication number Publication date
CN115914086A (en) 2023-04-04

Similar Documents

Publication Publication Date Title
EP3039817B1 (en) Determination and use of link performance measures
US10084665B1 (en) Resource selection using quality prediction
US7903573B2 (en) Method and system for network traffic matrix analysis
CN106165354B (en) Network path selection using bandwidth prediction
JP5051252B2 (en) Network failure detection system
US9602370B2 (en) Determining overall network health and stability
US20220360502A1 (en) Adaptive stress testing of sd-wan tunnels for what-if scenario model training
US20200389390A1 (en) Per-flow call admission control using a predictive model to estimate tunnel qos in sd-wan networks
CN113315682A (en) Method, system and apparatus for generating information transmission performance warning
CN115562879B (en) Computing power sensing method, computing power sensing device, electronic equipment and storage medium
Houidi et al. Constrained deep reinforcement learning for smart load balancing
WO2023016539A1 (en) Network path processing method and apparatus, storage medium, and electronic device
JP2019169783A (en) Routing control system, routing control method, and program
Barakabitze et al. QoEMultiSDN: Management of multimedia services using MPTCP/SR in softwarized and virtualized networks
Kumar et al. Routing path determination using QoS metrics and priority based evolutionary optimization
Lemeshko et al. Fast ReRoute model with multimedia quality protection
Mogyorósi et al. Adaptive protection of scientific backbone networks using machine learning
Watabe et al. A parallel flow monitoring technique that achieves accurate delay measurement
Nguyen et al. Evaluation of an end-to-end delay estimation in the case of multiple flows in SDN networks
Lai et al. A segment list selection algorithm based on delay in segment routing
Anh et al. A new localized multiconstraint QoS routing algorithm
US11894872B2 (en) Route selection in optical networks based on machine learning
Gunavathie et al. DLBA-A Dynamic Load-balancing Algorithm in Software-Defined Networking
Randhawa et al. NIMA (Network Impact Modeling and Analysis): A QoS Perspective
Hai-Anh et al. Towards a Novel Congestion Notification Algorithm for a Software-Defined Data Center Networks

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22855519

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE