CN116633864B - Flow scheduling method based on cloud computing platform - Google Patents
Flow scheduling method based on cloud computing platform Download PDFInfo
- Publication number
- CN116633864B CN116633864B CN202310887642.7A CN202310887642A CN116633864B CN 116633864 B CN116633864 B CN 116633864B CN 202310887642 A CN202310887642 A CN 202310887642A CN 116633864 B CN116633864 B CN 116633864B
- Authority
- CN
- China
- Prior art keywords
- traffic
- scheduling
- optimal
- flow
- link
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000005540 biological transmission Effects 0.000 claims abstract description 36
- 230000004044 response Effects 0.000 claims abstract description 18
- 239000002245 particle Substances 0.000 claims description 35
- 230000003044 adaptive effect Effects 0.000 claims description 18
- 238000004422 calculation algorithm Methods 0.000 claims description 14
- 230000006978 adaptation Effects 0.000 claims description 12
- 230000001133 acceleration Effects 0.000 claims description 5
- 238000012163 sequencing technique Methods 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 4
- 238000007619 statistical method Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005206 flow analysis Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
-
- 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
- H04L45/74—Address processing for routing
- H04L45/741—Routing in networks with a plurality of addressing schemes, e.g. with both IPv4 and IPv6
-
- 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
- H04L45/74—Address processing for routing
- H04L45/745—Address table lookup; Address filtering
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/25—Flow control; Congestion control with rate being modified by the source upon detecting a change of network conditions
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/29—Flow control; Congestion control using a combination of thresholds
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The application provides a traffic scheduling method based on a cloud computing platform, which comprises the steps of initially scheduling traffic based on a scheduling threshold; judging whether the flow is greater than an upper limit scheduling threshold, if so, sending a DNS query request, and sending a request message to a DNS cloud server; receiving a response message returned by the DNS cloud server, and checking a plurality of IP addresses of the DNS response; modeling a plurality of IP addresses, and outputting links corresponding to the optimal IP addresses according to the principle of priority of each flow rate corresponding to the plurality of IP addresses; and dynamically scheduling the traffic which is larger than the upper limit scheduling threshold value to a link corresponding to the optimal IP address, dynamically controlling the polling period, transmitting the traffic according to the transmission time length, and relatively uniformly distributing the traffic to the links which can be borne, thereby effectively relieving the congestion and reducing the problem of traffic blockage caused by unbalanced traffic distribution scheduling.
Description
Technical Field
The application relates to the technical field of telecommunications, in particular to a traffic scheduling method based on a cloud computing platform.
Background
With the increase of the traffic of each core part of the existing network, the access amount and the data traffic are rapidly increased, and the processing capacity and the computation intensity are correspondingly increased, so that a single server device cannot bear the huge load of links, and in order to share the link loads, a plurality of links are often provided in the network to bear the link loads;
the prior art CN104767690B discloses a traffic scheduling device and method, applied to an egress network device, the method comprising: when a DNS request message is received, acquiring a domain name type and a destination IP address in the DNS request message, wherein the destination IP address is a first IP address; searching a second IP address of a DNS server corresponding to the domain name type in a preset scheduling information table according to the domain name type, if the second IP address is different from the target IP address, modifying the target IP address of the DNS request message from the first IP address to the second IP address, and storing a modification record of the DNS request message in a session table; sending out the DNS request message according to the modified destination IP address;
there are the following problems:
the flexibility is poor, a second IP address is determined through a scheduling information table, and if the second IP address is different from the target IP address, the target IP address of the DNS request message is modified from the first IP address to the second IP address;
the network bandwidth which is faced to the increased flow can not change with the condition of accessing the flow at any time, and the surge of the flow at any time can cause serious congestion problem to the bandwidth, so that the experience of the user is poor.
Disclosure of Invention
In order to solve the problems, the application provides a flow scheduling method based on a cloud computing platform, so as to more exactly solve the problems of poor flexibility and uneven flow access;
there are the following problems:
the application is realized by the following technical scheme:
the application provides a traffic scheduling method based on a cloud computing platform, which comprises the following steps:
s1: carrying out initial scheduling on the traffic based on a scheduling threshold;
s2: the scheduling threshold is divided into an upper limit scheduling threshold and a lower limit scheduling threshold, whether the flow is larger than the upper limit scheduling threshold is judged, if so, a DNS query request is sent, and a request message is sent to a DNS cloud server;
s3: receiving a response message returned by the DNS cloud server, and checking a plurality of IP addresses of the DNS response;
s4: modeling a plurality of IP addresses, and outputting links corresponding to the optimal IP addresses according to the priority principle of each flow rate corresponding to the plurality of IP addresses;
s5: dynamically scheduling the traffic which is larger than the upper limit scheduling threshold value to the link corresponding to the optimal IP address, and dynamically controlling the polling period.
Further, in the flow scheduling method based on the cloud computing platform, the step S4 includes: inputting a plurality of IP addresses for modeling training, generating a plurality of particles in a feasible domain, combining an adaptation value of each particle with a flow rate of a network flow, and continuously and iteratively solving links corresponding to the optimal IP addresses; link representation，/>Representing a set of nodes>Representing edge set, ++>The starting point is +.>To the end->Is the link bandwidth representation->，/>The network stream is transmitted in the link, network stream +.>Is->，/>On-edge for k network flows>Flow distributed upwards, +.>Represents the kth flow vector,/->Representing the network flow k delivering a cost vector, the traffic rate of the network flow may be represented by:
，
the constraint conditions are that，/>，/>Representing the traffic rate of the network flow.
Further, in the flow scheduling method based on the cloud computing platform, the adaptive value updates the motion direction through 2 extremum, and the individual extremumAnd global extremum->；
The speed update formula is expressed as follows:
，
the location update formula is expressed as follows:
，
wherein the method comprises the steps ofRepresents the number of iterations, +.>Indicating that the particle is +.>Speed of time,/->The position of the particles is indicated,indicating particle->Is (are) located>Indicating that the particle is +.>Speed of->Acceleration weight representing the pushing of particles towards the optimal position of the individual,/->Acceleration weights representing the pushing of particles to the optimal position of the population; />Representing the inertial weighting factors,representing global extremum->Representing an even distribution in->The random number between the two, according to the merit of the particle of adaptive value calculation, the adaptive value objective function value can be confirmed by the following formula:
，
，
wherein the method comprises the steps ofRepresenting the fitness value of the objective function,/->Representing constraint violation values ∈>、/>The weighting coefficients of the equations are represented.
Further, in the flow scheduling method based on the cloud computing platform, the objective function adaptation value calculates an individual extremum and a global extremum of each particle; simultaneously performing multiple iterations, and stopping optimization when the difference between the adaptive value of the optimal solution after the previous iteration and the adaptive value of the optimal solution after the current iteration is smaller than a preset fitness threshold value;
if it isThen the optimal IP address for individual traffic is +.>If (if)Then there is global traffic corresponding to the optimal IP address +.>Wherein->Representation->Individual objective function adaptation value of +.>Representation->Individual objective function adaptation value of +.>Representation->Global objective function adaptation value of +.>Representing an individual preset fitness threshold, +.>Representing a global preset fitness threshold.
Further, in the flow scheduling method based on the cloud computing platform, the step of dynamically scheduling the flow greater than the upper limit scheduling threshold to the link corresponding to the optimal IP address includes: according to the scheduling algorithm, sequencing and dynamically scheduling the traffic transmission time to the link corresponding to the optimal IP address; updating the link flow queuing ready time until the time between the two links reaches less than the upper limit scheduling threshold;
ordered by size according to flowThe optimal IP address is allocated to the link corresponding to the optimal IP address:
，
，
，
wherein the method comprises the steps ofIndicating the flow output size, +.>Representing the flow input size, +.>Indicating the time required for traffic to be input to output in the link,/->Representing link traffic time,/->Representing the traffic volume,/->Representing link transmission capability, +.>Indicating the total time of link traffic.
Further, in the flow scheduling method based on the cloud computing platform, the step of dynamically scheduling the flow transmission time to the link corresponding to the optimal IP address according to the scheduling algorithm, the method includes:
calculating the time for completing the transmission flow of the link, selecting the shortest time, and distributing the shortest time to the link corresponding to the optimal IP address for transmission;
deleting tasks for scheduling traffic from the transmission tasks;
updating the time of other transmission traffic, and transmitting the shortest time traffic to the link corresponding to the optimal IP address again;
if the scheduling algorithm is smaller than the upper limit scheduling threshold, the scheduling algorithm is ended, and the program is exited.
Further, in the traffic scheduling method based on the cloud computing platform, the DNS cloud server performs balanced forwarding on each IP address by using a method of polling the selected interface based on the message as a unit.
Further, the flow scheduling method based on the cloud computing platform, before the step S1, includes: statistical analysis is performed on Ipv6 link traffic based on the sFlow collector.
Further, the flow scheduling method based on the cloud computing platform, after the step S2, includes: if the traffic scheduling threshold is smaller than the lower limit scheduling threshold, the downlink instruction stops the traffic scheduling.
The application has the beneficial effects that:
calculating the time for completing the transmission flow of the link according to the size of the transmission flow, selecting the shortest time, and distributing the shortest time to the link corresponding to the optimal IP address for transmission; deleting tasks for scheduling traffic from the transmission tasks; updating the time of other transmission traffic, and transmitting the shortest time traffic to the link corresponding to the optimal IP address again; if the scheduling algorithm is smaller than the upper limit scheduling threshold, ending the scheduling algorithm and exiting the program; the traffic is transmitted according to the transmission time, and the traffic is distributed to the links which can be borne in a relatively balanced way, so that the congestion is effectively relieved, the problem of traffic blockage caused by unbalanced traffic distribution and scheduling can be reduced, and the self-adaptability of the system is ensured; the flow scheduling is carried out based on the Ipv6 link, and the optional information of the IPv4 is replaced by the extension header according to the security mechanism of the Ipv6, so that the header of the Ipv6 is simplified, and the expansibility of the Ipv6 is enhanced; the length of a route in a routing table is reduced, the speed of forwarding a data packet by the router is improved, the speed in traffic scheduling based on an IPv6 link is faster than the scheduling speed of an IPV4, and the number of IP addresses is greatly improved;
(2) According to the application, a plurality of IP addresses are input for modeling training, a certain random generation of a plurality of particles is generated in a feasible domain, the adaptive value of each particle is combined with the flow rate of the network flow, the link corresponding to the optimal IP address is continuously and iteratively solved, the load sharing is carried out on the flow through the link corresponding to the optimal IP address, the load balancing is realized, the flexibility is better, the optimal IP addresses are compared with a plurality of IP addresses, and the flexible optimization is realized;
(3) The application provides a method for dynamically determining a polling period, which dynamically adjusts the length of the polling period according to the load of a network before polling, reduces the polling period when the load is light, and increases the polling period when the load is heavy, so that the problem of low utilization rate of an uplink channel when the load of the network is light is avoided, the probability of occurrence of the problem when the load of the network is heavy is also reduced, the polling period is dynamically controlled, the utilization rate of the uplink channel is improved, and the corresponding bandwidth is adjusted according to the flow scales of different server ends by optimizing an algorithm on the basis of the existing server scale, thereby enabling the whole system to operate in a high-efficiency state.
Drawings
FIG. 1 is a flow chart of a flow scheduling method based on a cloud computing platform;
fig. 2 is a schematic structural diagram of a flow dispatching system based on a cloud computing platform according to the present application.
Description of the embodiments
In order to more clearly and completely describe the technical scheme of the application, the application is further described below with reference to the accompanying drawings.
Referring to fig. 1, the present application proposes a flow chart of a flow scheduling method based on a cloud computing platform;
s1: and carrying out initial scheduling on the traffic based on the scheduling threshold.
In a specific embodiment, before initial scheduling is performed on traffic, measurement is required on Ipv6 link traffic; statistical analysis is carried out on the flow of the Ipv6 link based on an sFlow collector, and the sFlow provides flow analysis based on an interface, so that the flow condition can be monitored in real time, and the sources of abnormal flow and attack flow can be found in time; the sFlow system comprises an sFlow agent embedded in the device and a remote sFlow collector; the sFlowAgents acquire interface statistical information and data information through sFlow sampling, package the information into sFlow messages, and send the sFlow messages to a designated sFlowCollector when an sFlow message buffer is full or after the time of the sFlow message buffer is overtime (the buffer time is 1 second); the sFlow collector analyzes the sFlow message, displays an analysis result, and performs initial scheduling on the flow analyzed by the sFlow based on a scheduling threshold, so that the number of IP addresses is increased, an IPv6 link has a safety mechanism based on the IPv6, the optional information of the IPv4 is replaced by an extension header, the header of the IPv6 is simplified, and the expansibility of the IPv6 is enhanced; the length of the route in the routing table is reduced, the speed of forwarding the data packet by the router is improved, the speed in the traffic scheduling based on the IPv6 link is faster than the scheduling speed of the IPV4, and the number of IP addresses is greatly improved.
S2: the scheduling threshold is divided into an upper limit scheduling threshold and a lower limit scheduling threshold, whether the flow is larger than the upper limit scheduling threshold is judged, if so, a DNS query request is sent, and a request message is sent to a DNS cloud server.
In specific implementation, the scheduling threshold is divided into an upper limit scheduling threshold and a lower limit scheduling threshold, wherein the upper limit scheduling threshold is set to 60% of the link traffic bearing, the lower limit scheduling threshold is set to 30% of the link traffic bearing, and if the lower limit scheduling threshold is smaller than the lower limit scheduling threshold, a stop instruction is issued to stop traffic scheduling; if the traffic is smaller than the upper limit scheduling threshold and larger than the lower limit scheduling threshold, the link traffic is indicated to be not congested, at the moment, the traffic feedback correction can be carried out, the route can be selected through ISP route, the user traffic can be uniformly distributed to different links for transmission on the premise of not being congested, and the occurrence of the condition that the traffic is too congested in the links is reduced; if the request message is greater than the upper limit scheduling threshold, a DNS request is sent to a DNS cloud server, the DNS cloud server detects the source of the message, if the message is found to be sent from the same link, the DNS server preferentially inquires whether the IP addresses corresponding to the same link are responded or not, if so, whether other links on the IP addresses are greater than the upper limit threshold scheduling is detected, if not, whether the IP addresses of other links are responded or not is inquired, or the DNS server does not support static proximity, and all the responded IP addresses recorded by the DNS server and the corresponding links are directly returned.
S3: and receiving a response message returned by the DNS cloud server, and checking a plurality of IP addresses of the DNS response.
In the implementation, a response IP address returned by the DNS server and a corresponding link are received, and whether a plurality of IP addresses of the DNS response are valid or not is detected; if the DNS response result is one IP address, changing the IP address to be the optimal IP address capable of carrying out traffic scheduling, and if the DNS response result is a plurality of IP addresses, modifying the received DNS response result: and sequencing the plurality of IP addresses according to the current flow rates of the interfaces corresponding to the links, which are acquired by the gateway equipment in real time, wherein the link with the highest current flow rate is used as the access priority sequencing.
S4: modeling a plurality of IP addresses, and outputting links corresponding to the optimal IP addresses according to the principle that each flow rate corresponding to the IP addresses is prioritized during measurement.
In the specific implementation, a plurality of IP addresses which are prioritized according to the rate are input for modeling training, a certain number of particles are generated in a feasible domain at random, the adaptive value of each particle is combined with the flow rate of the network flow, and links corresponding to the optimal IP addresses are continuously and iteratively solved; if the traffic of the original link is too much, the link corresponding to the optimal IP address cannot be loaded, selecting a second optimal IP address for common loading according to the same method; congested Link representation，/>Representing a set of nodes>Representing edge set, ++>The starting point is +.>To the end->Is the link bandwidth representation->，/>The network stream is transmitted in the link, network stream +.>Is->，/>On-edge for k network flows>Flow distributed upwards, +.>Represents the kth flow vector,/->Representing the network flow k delivering a cost vector, the traffic rate of the network flow may be represented by:
，
the constraint conditions are that，/>；
Generating randomly generated particles in a feasible domain, initializing the positions and the speeds of the particles, calculating an adaptive value for each particle, and then solving an optimal IP corresponding link through continuous iteration; each particle updates the direction of motion by 2 extrema, one being an individual extremumI.e. the optimal solution for the current flow, the other being the global extremum +.>I.e., the optimal solution of the network flow;
the speed update formula is expressed as follows:
，
the location update formula is expressed as follows:
，
wherein the method comprises the steps ofRepresents the number of iterations, +.>Indicating that the particle is +.>Speed of time,/->The position of the particles is indicated,indicating particle->Is (are) located>Indicating that the particle is +.>Speed of->Acceleration weight representing the pushing of particles towards the optimal position of the individual,/->Acceleration weights representing the pushing of particles to the optimal position of the population; />Representing the inertial weighting factors,representing global extremum->Representing an even distribution in->The random number between the two, according to the merit of the particle of adaptive value calculation, the adaptive value objective function value can be confirmed by the following formula:
，
，
wherein the method comprises the steps ofRepresenting the fitness value of the objective function,/->Representing constraint violation values ∈>、/>The weighting coefficients representing the equations;
calculating an individual extremum and a global extremum of each particle by using the adaptive value of the objective function; simultaneously performing multiple iterations, and stopping optimization when the difference between the adaptive value of the optimal solution after the previous iteration and the adaptive value of the optimal solution after the current iteration is smaller than a preset fitness threshold value;
if it isThen the optimal IP address for individual traffic is +.>If (if)Then there is global traffic corresponding to the optimal IP address +.>Wherein->Representation->Individual objective function adaptation value of +.>Representation->Individual objective function adaptation value of +.>Representation->Global objective function adaptation value of +.>The preset fitness threshold value of the individual is represented, the optimal 50% of the individual flow can be represented,representing a global preset fitness threshold; />The method comprises the steps of representing a global preset fitness threshold value, namely, the optimal 80% of global particle flow, and collecting the state of a network and the related information of link congestion through the sampling capability of sFlow; then, according to the response IP of the DNS calculation server, a feasible path is obtained, and the diversity of links is increased; and finally, modeling the multi-path selection of the feasible IP based on the multi-commodity flow problem, and solving the optimal solution by using a particle swarm algorithm, so that the congestion link can avoid the congestion relieving path, the overall utilization rate of the network is improved, if the optimal IP corresponds to a plurality of links, the plurality of links on the IP are compared in terms of flow utilization rate, and if the utilization rate is lower and higher, the link corresponds to the optimal IP.
S5: dynamically scheduling the traffic which is larger than the upper limit scheduling threshold value to the link corresponding to the optimal IP address, and dynamically controlling the polling period.
In the specific steps, dynamically scheduling the traffic which is larger than an upper limit threshold value to a link corresponding to an optimal IP address according to a scheduling algorithm; sequencing the traffic transmission time, scheduling the traffic to the link corresponding to the optimal IP address according to the traffic with the priority of short transmission time, and calculating the traffic transmission time by combining the traffic rate of the link corresponding to the optimal IP address until the traffic queuing ready time of the link is updated when the traffic transmission time between the two links is smaller than the upper limit scheduling threshold;
ordered by size according to flowThe optimal IP address is allocated to the link corresponding to the optimal IP address:
，
，
，
wherein the method comprises the steps ofIndicating the flow output size, +.>Representing the flow input size, +.>Indicating the time required for traffic to be input to output in the link,/->Representing the bandwidth rate of the link corresponding to the optimal IP address, < >>Representing link traffic time,/->Representing the traffic volume,/->Representing link transmission capability, +.>Representing the total time of link traffic;
calculating the time for completing the link transmission flow according to the transmission flow and the link flow rate, selecting the shortest time, and distributing the shortest time to the link corresponding to the optimal IP address for transmission;
deleting tasks for scheduling traffic from the transmission tasks;
updating the time of other transmission traffic, and transmitting the shortest time traffic to the link corresponding to the optimal IP address again;
if the links corresponding to the original link and the optimal IP address are smaller than the upper limit scheduling threshold, the scheduling algorithm is ended, and the program is exited; if the condition that the link corresponding to the optimal IP address is larger than the upper limit scheduling threshold value occurs in the transmission process, if the condition lasts for 1min, the DNS cloud server needs to select the links corresponding to other IP addresses to help bear the load so as to avoid the condition of excessive link congestion;
the method comprises the steps of dynamically controlling a polling period, dynamically determining the polling period, dynamically adjusting the length of the polling period according to the load of a network before polling, shrinking the polling period when the load is light, and increasing the polling period when the load is heavy, so that the problem of low utilization rate of an uplink channel when the load is light is avoided, the probability of occurrence of the problem when the load is heavy is reduced, dynamically controlling the polling period, improving the utilization rate of the uplink channel, ensuring the bandwidth requirement when the data transmission of high-priority service is determined, starting to allocate the bandwidth from the medium-priority service when the data transmission of the high-priority service is determined, reducing the waste of the bandwidth, and improving the utilization rate of the bandwidth.
Referring to fig. 2, the present application proposes a schematic structural diagram of a flow scheduling system based on a cloud computing platform;
in one embodiment, an sFlow collector is arranged in the scheduling module, the sFlow collector collects and performs statistical analysis on the flow of an Ipv6 link, the sFlow provides flow analysis based on an Ipv6 interface, the flow condition can be monitored in real time, the source of abnormal flow and attack flow can be found in time, if the monitored flow of the link is larger than an upper limit threshold, the system starts to perform initial scheduling on the flow of link congestion based on the scheduling threshold, the scheduling module sends a DNS query request, a request message is sent to a DNS cloud server, the DNS cloud server performs balanced forwarding on each IP address by adopting a method based on a message as a unit, the DNS cloud server polls and selects an interface, the DNS cloud service can detect the source of the message, if the message is found to be sent from a first link, the DNS server preferentially queries whether the IP address corresponding to the first link is responded, if not responded, the IP addresses of other links are queried, and the response IP address returned by the DNS server and the corresponding link are received; the scheduling module receives whether a plurality of IP addresses of the DNS response are valid or not detected by the DNS cloud server; if the DNS response result is one IP address, the IP address is changed to be an optimal IP address capable of carrying out flow scheduling, if the DNS response result is a plurality of IP addresses, the plurality of IP addresses are ordered according to the current flow rates of interfaces corresponding to all links obtained by gateway equipment in real time, wherein a link with the highest current flow rate is used as access priority order, a plurality of addresses are modeled, an optimal IP address corresponding link is output according to the principle of priority of the flow rates, if the optimal IP corresponding link cannot be found, an optimal path is found again in the model, the flow which is larger than an upper limit threshold value is dynamically scheduled to the link corresponding to the optimal IP, the training period is dynamically controlled, and meanwhile, the IP address responded by the DNS cloud server and the content scheduling of the scheduling module are recorded in a storage module.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, apparatus, article or method that comprises the element.
The foregoing description is only of the preferred embodiments of the present application and is not intended to limit the scope of the application, and all equivalent structures or equivalent processes using the descriptions and drawings of the present application or direct or indirect application in other related technical fields are included in the scope of the present application.
Although embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the application, the scope of which is defined in the appended claims and their equivalents.
Of course, the present application can be implemented in various other embodiments, and based on this embodiment, those skilled in the art can obtain other embodiments without any inventive effort, which fall within the scope of the present application.
Claims (6)
1. The traffic scheduling method based on the cloud computing platform is characterized by comprising the following steps of:
s1: carrying out initial scheduling on the traffic based on a scheduling threshold;
s2: the scheduling threshold is divided into an upper limit scheduling threshold and a lower limit scheduling threshold, whether the flow is larger than the upper limit scheduling threshold is judged, if so, a DNS query request is sent, and a request message is sent to a DNS cloud server;
s3: receiving a response message returned by the DNS cloud server, and checking a plurality of IP addresses of the DNS response;
s4: modeling a plurality of IP addresses, and outputting links corresponding to the optimal IP addresses according to the priority principle of each flow rate corresponding to the plurality of IP addresses;
s5: dynamically scheduling the traffic which is larger than the upper limit scheduling threshold value to a link corresponding to the optimal IP address, and dynamically controlling the polling period;
the step S4 includes: inputting a plurality of IP addresses for modeling training, generating a plurality of particles in a feasible domain, combining an adaptation value of each particle with a flow rate of a network flow, and continuously and iteratively solving links corresponding to the optimal IP addresses; link representation，/>Representing a set of nodes>Representing edge set, ++>The starting point is +.>To the end->Is the link bandwidth representation->，/>The network flows are transmitted in links, and the bandwidth requirement of K network flows is d K ，/>On-edge for the kth network flow +.>Flow distributed upwards, +.>Represents the kth flow vector,/->Representing a kth network flow delivery cost vector, the traffic rate of the network flow may be represented by:
，
the constraint conditions are that，/>，/>Representing the traffic rate of the kth network flow;
the adaptive value updates the motion direction through 2 extreme values, and the individual extreme valuesAnd global extremum->；
The speed update formula is expressed as follows:
，
the location update formula is expressed as follows:
，
wherein the method comprises the steps ofRepresents the number of iterations, +.>Indicating that the particle is +.>Speed of time,/->Indicating the position of the particle, +.>Indicating particle->Is (are) located>Indicating that the particle is +.>Speed of->Indicating that particles are to beAcceleration weight pushing to the optimal position of the individual, +.>Acceleration weights representing the pushing of particles to the optimal position of the population; />Representing inertial weighting factors, +.>Representing global extremum->Representing an even distribution in->The random number between the two, according to the merit of the particle of adaptive value calculation, the adaptive value objective function value can be confirmed by the following formula:
，
，
wherein the method comprises the steps ofRepresenting the fitness value of the objective function,/->Representing constraint violation values ∈>、/>The weighting coefficients representing the equations;
the objective function adaptation value calculates an individual extremum and a global extremum of each particle; simultaneously performing multiple iterations, and stopping optimization when the difference between the adaptive value of the optimal solution after the previous iteration and the adaptive value of the optimal solution after the current iteration is smaller than a preset fitness threshold value;
if it isThen the optimal IP address for individual traffic is +.>If (if)Then there is global traffic corresponding to the optimal IP address +.>WhereinRepresentation->Individual objective function adaptation value of +.>Representation->Is used to adapt the individual objective function of the (c) to the value,representation->Global objective function adaptation value of +.>Representing an individual preset fitness threshold, +.>Representing a global preset fitness threshold.
2. The traffic scheduling method based on the cloud computing platform according to claim 1, wherein the step of dynamically scheduling traffic greater than an upper limit scheduling threshold to a link corresponding to an optimal IP address comprises: according to the scheduling algorithm, sequencing and dynamically scheduling the traffic transmission time to the link corresponding to the optimal IP address; updating the link flow queuing ready time until the time between the two links reaches less than the upper limit scheduling threshold;
ordered by size according to flowThe optimal IP address is allocated to the link corresponding to the optimal IP address:
，
，
，
wherein the method comprises the steps ofIndicating the flow output size, +.>Representing the flow input size, +.>Indicating the time required for traffic to be input to output in the link,/->Indicating the time of the link traffic transmission,/>representing the traffic volume,/->Representing link transmission capability, +.>Indicating the total time of link traffic.
3. The method for traffic scheduling based on the cloud computing platform according to claim 2, wherein the step of dynamically scheduling the traffic transmission time according to the scheduling algorithm to the link corresponding to the optimal IP address includes:
calculating the time for completing the transmission flow of the link, selecting the shortest time, and distributing the shortest time to the link corresponding to the optimal IP address for transmission;
deleting tasks for scheduling traffic from the transmission tasks;
updating the time of other transmission traffic, and transmitting the shortest time traffic to the link corresponding to the optimal IP address again;
if the scheduling algorithm is smaller than the upper limit scheduling threshold, the scheduling algorithm is ended, and the program is exited.
4. The traffic scheduling method based on the cloud computing platform as claimed in claim 1, wherein the DNS cloud server performs balanced forwarding on each IP address by polling the selection interface based on the unit of the message.
5. The flow scheduling method based on the cloud computing platform according to claim 1, wherein before the step of S1, the method comprises: statistical analysis is performed on Ipv6 link traffic based on the sFlow collector.
6. The flow scheduling method based on the cloud computing platform as claimed in claim 1, wherein after the step S2, the method comprises: if the traffic scheduling threshold is smaller than the lower limit scheduling threshold, the downlink instruction stops the traffic scheduling.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310887642.7A CN116633864B (en) | 2023-07-19 | 2023-07-19 | Flow scheduling method based on cloud computing platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310887642.7A CN116633864B (en) | 2023-07-19 | 2023-07-19 | Flow scheduling method based on cloud computing platform |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116633864A CN116633864A (en) | 2023-08-22 |
| CN116633864B true CN116633864B (en) | 2023-11-03 |
Family
ID=87642080
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310887642.7A Active CN116633864B (en) | 2023-07-19 | 2023-07-19 | Flow scheduling method based on cloud computing platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116633864B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102572001A (en) * | 2011-01-04 | 2012-07-11 | 中兴通讯股份有限公司 | Domain name system (DNS) and method for providing load balancing |
| CN104767690A (en) * | 2014-01-08 | 2015-07-08 | 杭州迪普科技有限公司 | Flow scheduling device and method |
| CN106992938A (en) * | 2017-05-15 | 2017-07-28 | 网宿科技股份有限公司 | A kind of network traffics dynamic dispatching distribution method and system |
| CN110891019A (en) * | 2019-12-10 | 2020-03-17 | 重庆邮电大学 | Data center flow scheduling method based on load balancing |
| US10778757B1 (en) * | 2018-06-18 | 2020-09-15 | Amazon Technologies, Inc. | Load balancing traffic via dynamic DNS record TTLs |
| WO2022267791A1 (en) * | 2021-06-22 | 2022-12-29 | 南京邮电大学 | Workflow scheduling method and system based on multi-objective particle swarm algorithm, and storage medium |
| CN115835300A (en) * | 2022-11-03 | 2023-03-21 | 互联网域名***北京市工程研究中心有限公司 | DNS-based traffic balance scheduling method and device and gateway equipment |
-
2023
- 2023-07-19 CN CN202310887642.7A patent/CN116633864B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102572001A (en) * | 2011-01-04 | 2012-07-11 | 中兴通讯股份有限公司 | Domain name system (DNS) and method for providing load balancing |
| CN104767690A (en) * | 2014-01-08 | 2015-07-08 | 杭州迪普科技有限公司 | Flow scheduling device and method |
| CN106992938A (en) * | 2017-05-15 | 2017-07-28 | 网宿科技股份有限公司 | A kind of network traffics dynamic dispatching distribution method and system |
| US10778757B1 (en) * | 2018-06-18 | 2020-09-15 | Amazon Technologies, Inc. | Load balancing traffic via dynamic DNS record TTLs |
| CN110891019A (en) * | 2019-12-10 | 2020-03-17 | 重庆邮电大学 | Data center flow scheduling method based on load balancing |
| WO2022267791A1 (en) * | 2021-06-22 | 2022-12-29 | 南京邮电大学 | Workflow scheduling method and system based on multi-objective particle swarm algorithm, and storage medium |
| CN115835300A (en) * | 2022-11-03 | 2023-03-21 | 互联网域名***北京市工程研究中心有限公司 | DNS-based traffic balance scheduling method and device and gateway equipment |
Non-Patent Citations (3)
| Title |
|---|
| 《基于粒子群优化算法的数据中心网络流量调度策略》;马枢清等;《电讯技术》;第61卷(第7期);全文 * |
| 基于智能DNS的网络流量负载均衡控制研究;李雨泰;来风刚;尚智婕;董希杰;;自动化与仪器仪表(第11期);全文 * |
| 李雨泰 ; 来风刚 ; 尚智婕 ; 董希杰 ; .基于智能DNS的网络流量负载均衡控制研究.自动化与仪器仪表.2018,(第11期),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116633864A (en) | 2023-08-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Prabhavat et al. | Effective delay-controlled load distribution over multipath networks | |
| De Veciana et al. | Stability and performance analysis of networks supporting elastic services | |
| CN102075444B (en) | Network system and method for guaranteeing multi-type service quality | |
| EP1035703B1 (en) | Method and apparatus for load sharing on a wide area network | |
| CN109714268B (en) | Flow control method and related device for virtual private cloud | |
| US7646709B2 (en) | Flow control in computer networks | |
| US9042355B2 (en) | Quality of service (QoS) for satellite communications network | |
| Banaie et al. | Load-balancing algorithm for multiple gateways in Fog-based Internet of Things | |
| EP3883210B1 (en) | Method and apparatus for determining forwarding port in information centeric network | |
| JP5888687B1 (en) | Distribution server or distribution route design device, distribution server or distribution route design method and program | |
| US9806987B2 (en) | Method and apparatus for reducing response time in information-centric networks | |
| CN110650099B (en) | Queuing system for predicting packet lifetime in a computing device | |
| Kim et al. | Differentiated forwarding and caching in named-data networking | |
| CN116909735A (en) | Calculation power scheduling method and device, server and storage medium | |
| EP3278500B1 (en) | Processing data items in a communications network | |
| Vladyko et al. | Fuzzy model of dynamic traffic management in software-defined mobile networks | |
| JP2018098590A (en) | Traffic demand prediction device, traffic demand prediction method, and program | |
| CN116633864B (en) | Flow scheduling method based on cloud computing platform | |
| KR101465884B1 (en) | Method and apparatus of probabilistic controller selection in software-defined networks | |
| CN103905327A (en) | Stream state information based content-centric network congestion control method and system thereof | |
| Pan et al. | SHRiNK: a method for enabling scaleable performance prediction and efficient network simulation | |
| CN107231318A (en) | Switching at runtime control to providing multiple Internet services for home network | |
| Nepolo et al. | A predictive ECMP routing protocol for fat-tree enabled data centre networks | |
| Ye et al. | Dynamic optimization of OSPF weights using online simulation | |
| Alali et al. | Calibers: A bandwidth calendaring paradigm for science workflows |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |