WO2022116665A1 - Tcp流的调整方法和*** - Google Patents

Tcp流的调整方法和*** Download PDF

Info

Publication number
WO2022116665A1
WO2022116665A1 PCT/CN2021/121789 CN2021121789W WO2022116665A1 WO 2022116665 A1 WO2022116665 A1 WO 2022116665A1 CN 2021121789 W CN2021121789 W CN 2021121789W WO 2022116665 A1 WO2022116665 A1 WO 2022116665A1
Authority
WO
WIPO (PCT)
Prior art keywords
tcp flow
tcp
address
data
destination
Prior art date
Application number
PCT/CN2021/121789
Other languages
English (en)
French (fr)
Inventor
江舟
连超
赵军锋
张平荣
Original Assignee
中兴通讯股份有限公司
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 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2022116665A1 publication Critical patent/WO2022116665A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/10—Network architectures or network communication protocols for network security for controlling access to devices or network resources
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
    • H04L69/16—Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
    • H04L69/161—Implementation details of TCP/IP or UDP/IP stack architecture; Specification of modified or new header fields
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
    • H04L69/16—Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
    • H04L69/163—In-band adaptation of TCP data exchange; In-band control procedures
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W80/00—Wireless network protocols or protocol adaptations to wireless operation
    • H04W80/06—Transport layer protocols, e.g. TCP [Transport Control Protocol] over wireless

Definitions

  • the embodiments of the present application relate to the field of communications, and in particular, to a method and system for adjusting a TCP flow.
  • 5G communication network has the characteristics of high speed (enhanced mobile broadband, eMBB), large capacity (large-scale machine communication, mMTC), and low latency (high reliability and low latency communication, uRLLC), so with the development of 5G communication network , greatly improving the user's network access experience.
  • eMBB enhanced mobile broadband
  • mMTC large-scale machine communication
  • uRLLC low latency
  • the embodiment of the present application provides a method for adjusting a TCP flow, which is applied to NWDAF (Network Data Analytics Function, network data analysis function), and the method includes:
  • SMF Session management function, respectively to session management function
  • UPF User Plane Function, user plane function
  • the SMF data and the UDF data are analyzed to estimate the TCP flow adjustment strategy
  • the embodiment of the present application also provides a method for adjusting a TCP flow, which is applied to a TCP acceleration device, and the TCP acceleration device is respectively connected to the network data analysis function NWDAF, the user plane function UPF and the external gateway in communication and connection, and the method includes:
  • the NWDAF receiving the TCP flow adjustment policy fed back by the NWDAF according to the TCP flow adjustment policy subscription instruction, where the TCP flow adjustment policy is determined by the NWDAF based on the SMF data fed back by the session management function SMF and the UPF data fed back by the UPF;
  • the TCP flow generated when each application program APP accesses the network is adjusted.
  • the embodiment of the present application also provides a device for adjusting a TCP flow, including:
  • the log subscription module is used to send log subscription instructions to the session management function SMF and the user plane function UPF respectively;
  • a data receiving module configured to receive the SMF data fed back by the SMF according to the log subscription instruction, and the UDF data fed back by the UPF according to the log subscription instruction;
  • a data analysis module for analyzing the SMF data and the UDF data, and estimating the TCP flow adjustment strategy
  • a policy push module configured to push the TCP flow adjustment policy to the TCP acceleration device pre-accessed between the UPF and the external gateway, so that the TCP acceleration device can access each application APP according to the TCP flow adjustment policy
  • the embodiment of the present application also provides a device for adjusting a TCP flow, including:
  • a policy subscription module for sending a TCP stream adjustment policy subscription instruction to the NWDAF
  • a policy receiving module configured to receive the TCP flow adjustment policy fed back by the NWDAF according to the TCP flow adjustment policy subscription instruction, where the TCP flow adjustment policy is fed back by the NWDAF based on the SMF data fed back by the session management function SMF and the UPF feedback The UPF data is determined;
  • the TCP flow adjustment module is configured to adjust the TCP flow generated when each application program APP accesses the network according to the TCP flow adjustment policy.
  • the embodiment of the present application also provides a network data analysis function, including:
  • the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the adjustment of the TCP stream applied to the NWDAF as described above method.
  • the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute the TCP stream applied to the TCP acceleration device as described above adjustment method.
  • Embodiments of the present application further provide a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, implements the above-mentioned adjustment method for a TCP stream applied to NWDAF, or is applied to TCP as described above A method for adjusting the TCP flow of the acceleration device.
  • the embodiment of the present application also provides a TCP flow adjustment system, including:
  • the NWDAF is also communicatively connected to the UPF;
  • the NWDAF is used to perform the adjustment method applied to the TCP flow of the NWDAF as described above;
  • the TCP acceleration device is configured to perform the adjustment method of the TCP flow applied to the TCP acceleration device as described above.
  • FIG. 1 is a schematic structural diagram of a TCP flow adjustment system provided by a first embodiment of the present application
  • FIG. 2 is a schematic diagram of the architecture of the TCP flow adjustment system provided in the first embodiment of the present application in a 5G network;
  • FIG. 3 is a flowchart of a method for adjusting a TCP flow provided by a second embodiment of the present application.
  • FIG. 4 is a schematic diagram of interaction of various functional entities involved in a method for adjusting a TCP flow provided by the second embodiment of the present application;
  • FIG. 5 is a flowchart of a method for adjusting a TCP flow provided by a third embodiment of the present application.
  • FIG. 6 is a schematic structural diagram of an apparatus for adjusting a TCP stream provided by a fourth embodiment of the present application.
  • FIG. 7 is a schematic structural diagram of an apparatus for adjusting a TCP flow provided by a fifth embodiment of the present application.
  • FIG. 8 is a schematic structural diagram of a network data analysis function provided by the sixth embodiment of the present application.
  • FIG. 9 is a schematic structural diagram of a TCP acceleration apparatus provided by a seventh embodiment of the present application.
  • the embodiments of the present application provide a method and system for adjusting a TCP (Transmission Control Protocol, Transmission Control Protocol) flow.
  • NWDAF Network Data Analytics Function
  • SMF Session management function
  • UPF User Plane Function
  • a TCP flow adjustment system includes: a TCP acceleration device 10 , a network data analysis function 20 , a session management function 30 , a user plane function 40 and an external gateway 50 .
  • the TCP acceleration device 10 is respectively connected to the network data analysis function 20 , the user plane function 40 and the external gateway 50 , and the network data analysis function 20 is in turn connected to the session management function 30 and the user plane function 40 .
  • the network data analysis function 20 mentioned in this embodiment is the NWDAF (Network Data Analytics Function) mentioned in the communication field. It is a functional entity proposed by the fifth-generation mobile communication technology (5th-Generation, 5G) system. It can be regarded as a data-aware analysis network element. It mainly uses network data as the basis to automatically perceive and analyze the network, and participate in In the whole life cycle of network planning, construction, operation and maintenance, network optimization, and operation, the network is easy to maintain and control, improve the efficiency of network resource use, and improve user service experience.
  • NWDAF Network Data Analytics Function
  • the session management function 30 is also a functional entity of the 5G service-based architecture, which is called SMF (Session management function) in the communication field.
  • SMF Session management function
  • the SMF is mainly responsible for interacting with the classified data plane, creating, updating and deleting protocol data unit (Protocol Data Unit, PDU) sessions, and managing the session context (session context) with the user plane function 40, That is, in practical applications, there is also a communication connection between the session management function 30 and the user plane function 40 .
  • the user plane function 40 is also an important functional entity in the 5G-based service architecture, which is the so-called UPF (User Plane Function) in the communication field.
  • UPF User Plane Function
  • the TCP acceleration device 10 is a functional entity newly added to the 5G service-based architecture in this embodiment, and is mainly used to accelerate or adjust the speed limit of the TCP flow according to the TCP flow adjustment policy provided by the network data sub-function 20 .
  • NSSF Network Slice Selection Function, network slice selection function
  • AUSF Authentication Server Function, authentication service function
  • NEF Network Exposure Function, network service presentation function
  • UDM Unified Data Management, unified Data management function
  • NRF Network Exposure Function, network service presentation function
  • UDM Unified Data Management, unified Data management function
  • NRF Network Exposure Function, network element data warehouse function
  • PCE Policy Control function, policy control function
  • AF Application Function, application layer function
  • AMF Access and Mobility Management Function, access and mobility management function
  • SMF Session management function, session management function
  • NWDAF Network Data Analytics Function, network data analysis function
  • the DN mentioned above is the external gateway mentioned in this embodiment.
  • the external gateway can be not only a DN, but also a FW (firewall, firewall), or a core network other network devices that carry the role of a gateway.
  • the above is only a specific architecture of the 5G network at present, and does not constitute a limitation on the technical solutions of the present application.
  • the specific network element functional entities in the architecture can be based on business needs. There are changes, as long as the architecture includes the TCP acceleration device 10, the network data analysis function 20, the session management function 30, the user plane function 40 and the external gateway 50 mentioned in this embodiment, and ensure that these five network element functional entities are Between them, the communication connection can be performed according to the connection relationship mentioned in this embodiment.
  • a TCP flow adjustment policy subscription instruction is sent by the TCPO (ie, the TCP acceleration device 10 in FIG. 1 ) to the NWDAF (ie, the network data analysis function 20 in FIG. 1 ).
  • the NWDAF After receiving the TCP flow adjustment policy subscription instruction sent by the TCPO, the NWDAF generates a log subscription instruction, and sends the generated log subscription instruction to the SMF (that is, the session management function 30 in FIG. 1 ) and the UPF (that is, the session management function 30 in FIG. 1 ).
  • the user plane function 40 in FIG. 1 so that the SMF can push the generated data (hereinafter referred to as SMF data for convenience of description) to the NWDAF when monitoring the creation, or update, or release of the PDU session;
  • the UF can push the generated data (hereinafter referred to as UPF data for convenience of description) to the NWDAF when the flow release operation is monitored.
  • NWDAF analyzes SMF data and UDF data, and then predicts a TCP flow adjustment policy suitable for the current network, and pushes the obtained TCP flow adjustment policy to TCPO.
  • TCPO adjusts the TCP flow generated when each APP (Application, application) released by UPF accesses the network, such as accelerating the TCP flow corresponding to the high-value APP, and then sending it to The external gateway DN, which limits the rate of the TCP flow corresponding to the low-value APP, and then sends it to the external gateway DN.
  • APP Application, application
  • the TCP flow adjustment system analyzes the SMF data provided by the SMF and the UPF data provided by the UPF by using the intelligent analysis statistics of the NWDAF and the trend prediction, and then predicts the access network of each APP.
  • the adjustment strategy of the TCP flow generated at the time such as accelerating the TCP flow of a certain type of APP, and limiting the speed of a certain type, and finally push the estimated TCP flow adjustment strategy to the TCP between the pre-connected UPF and the external gateway.
  • the acceleration device enables the TCP acceleration device to reasonably adjust the TCP flow from the UPF, thereby improving the user network access experience of the mobile terminal as much as possible under limited network resources.
  • the second embodiment of the present application relates to a method for adjusting a TCP flow, which is applied to the network data analysis function NWDAF.
  • Step 301 Send log subscription instructions to the session management function SMF and the user plane function UPF respectively.
  • Step 302 Receive the SMF data fed back by the SMF according to the log subscription instruction, and the UDF data fed back by the UPF according to the log subscription instruction.
  • Step 303 Analyze the SMF data and the UDF data to estimate a TCP flow adjustment policy.
  • Step 304 Push the TCP flow adjustment policy to the TCP acceleration device pre-connected between the UPF and the external gateway, for the TCP acceleration device to generate when each application APP accesses the network according to the TCP flow adjustment policy.
  • the TCP stream is adjusted.
  • TCPO TCP acceleration device
  • the premise of triggering the NWDAF to perform the operations from steps 301 to 304 is that the NWDAF has received the TCP flow adjustment policy subscription instruction sent by the TCPO (TCP acceleration device).
  • NWDAF After NWDAF receives the TCP stream adjustment policy subscription instruction sent by TCPO, it will generate log subscription instructions, and send the generated log subscription instructions to SMF and UPF respectively.
  • the log subscription command in practical applications, it specifically refers to the command to subscribe to the call history record log (call history record log, CHR log), that is, the subsequently received SMF data fed back by the SMF according to the log subscription command and UPF extracted from the CHR log when the UPF data fed back by the log subscription instruction.
  • call history record log call history record log, CHR log
  • the SMF is mainly responsible for interacting with the classified data plane, creating, updating and deleting PDU sessions. Therefore, when there is a PDU session generated by interactive operations such as creation, update, or deletion, the SMF will perform the operation of collecting information on the creation of the control plane, and push the collected data (hereinafter referred to as SMF data) to the NWDAF.
  • SMF data collected data
  • the UPF is mainly responsible for sending the UE, in this embodiment, it mainly refers to the flow generated when the APP in the mobile user equipment accesses the network, and in this embodiment, it mainly refers to sending the TCP flow to the external gateway, that is, the flow is released. Therefore, when a flow is released, the UPF will collect flow information, and push the collected data (hereinafter referred to as UPF data) to the NWDAF.
  • the NWDAF After receiving the SMF data and the UPF data, the NWDAF will analyze the SMF data and the UDF data based on its own analysis and prediction capabilities, and then estimate the TCP flow adjustment policy.
  • the NWDAF analyzes the SMF data and the UDF data, and then estimates the operation of obtaining the TCP flow adjustment policy. In practical applications, it is specifically:
  • the SMF data and the UDF data are combined according to preset field information to obtain data to be processed.
  • the SEID Near Field Communication NFC serial number
  • the N4 interface are used as preset field information, and the SMF data and the UDF data are combined, that is, the data provided by the N4 interface with the same SEID is combined.
  • the merged data to be processed mainly includes: International Mobile Subscriber Identity (IMSI), private network address, private network port, protocol type, destination address, destination port, number of packets, duration , known APP types, uniform resource locator (URL), etc.
  • IMSI International Mobile Subscriber Identity
  • private network address private network port
  • protocol type private network port
  • destination address destination address
  • destination port number of packets
  • duration time
  • known APP types uniform resource locator (URL), etc.
  • URL uniform resource locator
  • the TCP flow adjustment policy is estimated to be able to quickly identify APPs that occupy a large amount of traffic but have low value, as well as high-value APPs such as popular APPs.
  • the preset time granularity and the preset area are used as the dimensions, and the statistics of the different time dimensions and different usage areas in the data to be processed are calculated.
  • Source IP address, destination IP address corresponding to each APP in each client, and historical traffic usage information of each APP then, by traversing each APP, establish the source IP address and the traversed APP.
  • the mapping relationship between the destination IP address corresponding to the APP and the historical traffic usage information, the historical traffic usage statistics table of each APP is obtained; finally, based on the preset trend prediction machine learning algorithm, the historical traffic usage statistics table is obtained.
  • Each recorded piece of data with a mapping relationship is analyzed to obtain the TCP flow adjustment policy.
  • the unit may be minutes, hours, or days, that is, every minute, or an hour, or once a day; the preset area may be a cell as a unit.
  • the TCP acceleration device used to adjust the TCP flow can only identify the destination IP, and accelerate or limit the speed of the corresponding TCP flow through the destination IP. Therefore, the above-mentioned TCP flow adjustment policy specifically records the destination IP address that needs to be accelerated and the destination IP address that needs to be speed-limited.
  • NWDAF uses a preset trend prediction machine learning algorithm based on the analysis of each piece of data with a mapping relationship recorded in the historical traffic usage statistics table, and then identifies The destination IP address corresponding to the low-value APP that consumes a lot of traffic.
  • the low-value APPs that occupy a large amount of traffic in this embodiment are specifically obtained by NWDAF based on the data recorded in the historical traffic usage statistics table to be analyzed and obtained, and the top N of the network congestion phenomenon occur.
  • NWDAF the historical traffic usage of each cell, as well as the frequently occurring IMSIs of UEs and APPs that occupy a large amount of traffic but have a low proportion of usage when congestion occurs.
  • the UE is using a P2P-type APP to access the network, but it occupies a large amount of traffic allocated to the cell.
  • the P2P-type APPs used by the 0 UEs with different IMSIs are low-value APPs that occupy a large amount of traffic, that is, speed limiting is required.
  • the above-mentioned destination IP address that needs to be accelerated is specifically based on the preset trend prediction machine learning algorithm by NWDAF, by analyzing each piece of data with a mapping relationship recorded in the historical traffic usage statistics table, and then The destination IP address corresponding to the identified high-value APP.
  • high-value APPs in practical applications, it can be determined according to the popularity of users. For example, APPs with high user traffic can be considered as popular APPs. For this type of APP, it can be considered as a high-value APP.
  • the trend prediction machine learning algorithm analyzes the data in the historical traffic usage statistics table, it can take into account the impact of factors such as seasons and holidays on the trend, making the estimated TCP flow adjustment strategy more reasonable.
  • ARIMA Autoregressive Integrated Moving Average Model
  • LSTM Long Short-Term Memory Model
  • a network model with an LSTM and a fully connected layer can be pre-built, and then the constructed network model can be Perform iterative training to obtain a TCP flow adjustment strategy estimation model, so that each time new SMF data and UDP data are obtained, the two can be directly merged and input into the TCP flow adjustment strategy estimation model, and then the The output result of the TCP flow adjustment policy estimation model is used as the estimated TCP flow adjustment policy.
  • the pre-acquired sample data that is, the data obtained by merging the SMF data and UDP data generated at the historical moment, are divided into training set and test set according to the ratio of 7:3.
  • the flow values in the training set of the fixed time period are normalized, so as to improve the descending speed of the gradient term.
  • the normalized flow values are sequentially input to the constructed network model for iterative training.
  • the input flow values are the flow values at t-4, t-3, t-2, and t-1, respectively. Based on the flow values at these times, predict the flow value at time t, and then add the predicted flow value at time t to the training set to predict the flow value at time t+1.
  • the overall process is as follows:
  • the whole process iterates continuously until the granularity that needs to be predicted is reached.
  • the predicted value output after the above time series is input into the network model is compared with the actual value, the mean square error MSE loss function is used, and the ADAM optimization algorithm is used to optimize the loss function.
  • the loss function By optimizing the loss function, the gradient descent is continuously performed. After training, with the convergence of the loss function, a suitable network model can be obtained.
  • the corresponding traffic values at different times are input into the network model after iterative training, and then the output predicted value is compared with the actual real value.
  • the performance index of the test result reaches
  • the current network model can be determined as the TCP flow adjustment strategy prediction model, that is, the model can be used to predict a certain time in the future.
  • the traffic trend of the cell is entered, and a subsequent TCP flow adjustment policy suitable for the cell is formulated.
  • the growth rate of the upward trend exceeds the preset value, for example, the traffic increases by 30% within 30 minutes, the IMSI that causes the cell congestion will be counted from the historical data.
  • the corresponding private network IP and the destination IP accessed by the low-value APP are determined as the destination IP that needs to be speed-limited, and the TCP flow adjustment policy is pushed to the TCP acceleration device, and the TCP acceleration device performs the speed-limiting destination IP as required.
  • the TCP acceleration device will accelerate the target IP according to the needs, and the TCP flow generated when the APP corresponding to this purpose IP accesses the network will be processed. accelerate.
  • the TCPO after receiving the TCP flow adjustment policy pushed by NWDAF, the TCPO will identify the UE Internet access flow sent by UPF according to the destination IP address recorded in the TCP flow adjustment policy that needs to be accelerated. Then, accelerate the destination IP address that needs to be accelerated, that is, the TCP flow corresponding to the high-value APP, and then send the accelerated TCP flow to the external gateway; adjust the needs recorded in the policy according to the TCP flow
  • the destination IP address for speed limiting identify the destination IP address in the UE Internet access flow sent by UPF, and then limit the destination IP address that needs speed limiting, that is, the TCP flow corresponding to the low-value APP, and then limit the speed.
  • the subsequent TCP stream is sent to the external gateway.
  • the adjustment method of TCP flow uses the intelligent analysis statistics of NWDAF and trend prediction to analyze the SMF data provided by SMF and the UPF data provided by UPF, and then predict the correct The adjustment policy of the TCP flow generated when each APP accesses the network. For example, the TCP flow of a certain type of APP is accelerated, and the rate of a certain type is limited. Finally, the estimated TCP flow adjustment policy is pushed to the pre-connected UPF and external gateway. Therefore, the TCP acceleration device can reasonably adjust the TCP flow from the UPF, thereby improving the user network access experience of the mobile terminal as much as possible under limited network resources.
  • the TCP flow of the low-value APP is limited, and the high-value APP can be adjusted.
  • the third embodiment of the present application relates to a method for adjusting a TCP flow, which is applied to a TCP acceleration device.
  • TCP acceleration device other network element functional entities involved in implementing the TCP flow adjustment method, and the connection relationship with these network element functional entities, please refer to the TCP flow adjustment system provided by the first embodiment of this application for details. Corresponding Figures 1 and 2.
  • the adjustment method of the TCP flow involved in the second embodiment includes the following steps:
  • Step 501 Send a TCP flow adjustment policy subscription instruction to the NWDAF.
  • Step 502 Receive the TCP flow adjustment policy fed back by the NWDAF according to the TCP flow adjustment policy subscription instruction, and the TCP flow adjustment policy is based on the SMF data fed back by the NWDAF based on the session management function SMF and the UPF data fed back by the UPF. Sure.
  • Step 503 Adjust the TCP flow generated when each application program APP accesses the network according to the TCP flow adjustment policy.
  • the destination IP address that needs to be accelerated and the destination IP address that needs to be speed-limited are recorded in the TCP flow adjustment policy estimated by NWDAF.
  • the specific method is to read the destination IP address that needs to be accelerated and the destination IP address that needs to be speed-limited from the TCP flow adjustment policy. , and then, based on the destination IP address that needs to be accelerated, the TCP flow generated when the APP corresponding to the destination IP address that needs to be accelerated accesses the network is accelerated.
  • the TCP flow generated when the APP corresponding to the destination IP address that needs to be speed-limited accesses the network is speed-limited.
  • the estimated TCP flow adjustment policy since in practical application, NWDAF may perform analysis based on SMF data and UPF data, the estimated TCP flow adjustment policy only includes adjustment information for TCP flow, such as acceleration or speed limit.
  • the TCP acceleration device reads the destination IP address that needs to be accelerated and the destination IP address that needs to be speed-limited from the TCP flow adjustment policy, it can be judged first. Whether the destination IP address that needs to be accelerated and the destination IP address that needs to be speed-limited is carried in the TCP flow adjustment policy, and then a subsequent adjustment operation is triggered according to the judgment result.
  • the TCP flow call-out policy carries the destination IP address that needs to be accelerated and the destination IP address that needs to be speed-limited, then execute the acceleration according to the needs.
  • the destination IP address is to accelerate the TCP flow generated when the APP corresponding to the destination IP address that needs to be accelerated accesses the network, and the destination IP address that is speed-limited according to the needs, the speed-limited destination IP address that needs to be speed-limited.
  • the steps of speed limiting the TCP flow generated when the APP corresponding to the destination IP address accesses the network;
  • steps 501 to 503 in this embodiment are related to steps 301 to 304 in the second embodiment, that is, when adjusting the TCP flow, the implementation of the overall solution needs to involve step 301 Go to step 304, and all operations from step 501 to step 503.
  • the specific implementation of steps 501 to 503 involved in this embodiment has been described in the second embodiment in conjunction with the introduction of FIG. 4 , and details are not repeated here.
  • the method for adjusting the TCP flow analyzes the SMF data provided by the SMF and the UPF data provided by the UPF by using the intelligent analysis statistics of the NWDAF and the trend prediction, and then predicts the access network of each APP.
  • the adjustment strategy of the TCP flow generated at the time such as accelerating the TCP flow of a certain type of APP, and limiting the speed of a certain type, and finally push the estimated TCP flow adjustment strategy to the TCP between the pre-connected UPF and the external gateway.
  • the acceleration device enables the TCP acceleration device to reasonably adjust the TCP flow from the UPF, thereby improving the user network access experience of the mobile terminal as much as possible under limited network resources.
  • an apparatus for adjusting a TCP flow 600 includes a log subscription module 601 , a data receiving module 602 , a data analysis module 603 and a policy pushing module 604 .
  • the log subscription module 601 is configured to send log subscription instructions to the session management function SMF and the user plane function UPF respectively;
  • the data receiving module 602 is configured to receive the SMF data fed back by the SMF according to the log subscription instruction, and the UPF according to the UDF data fed back by the log subscription instruction;
  • the data analysis module 603 is used to analyze the SMF data and the UDF data, and estimate the TCP flow adjustment policy;
  • the policy push module 604 is used to The flow adjustment policy is pushed to the TCP acceleration device pre-connected between the UPF and the external gateway, for the TCP acceleration device to adjust the TCP flow generated when each application APP accesses the network according to the TCP flow adjustment policy.
  • the data analysis module 603 is specifically configured to combine the SMF data and the UDF data according to preset field information to obtain data to be processed; based on a preset trend prediction machine learning algorithm, to The preset time granularity and the preset area are dimensions, and the data in different time dimensions and different usage areas in the data to be processed are analyzed to obtain the TCP flow adjustment policy.
  • the preset trend prediction machine learning algorithm takes preset time granularity and preset area as dimensions, and performs data analysis on data in different time dimensions and different usage areas in the data to be processed. Analysis, to obtain the TCP flow adjustment strategy, specifically:
  • the preset trend prediction machine learning algorithm analyzes each piece of data with a mapping relationship recorded in the historical traffic usage statistics table to obtain the TCP flow adjustment policy, specifically for:
  • the TCP flow adjustment policy is obtained by analyzing the data.
  • the preset trend prediction machine learning algorithm analyzes each piece of data with a mapping relationship recorded in the historical traffic usage statistics table to obtain the TCP flow adjustment policy, specifically for:
  • the TCP flow adjustment policy is obtained.
  • the TCP flow adjustment policy is pushed to the TCP acceleration device that is pre-accessed between the UPF and the external gateway, so that the TCP acceleration device can adjust each flow according to the TCP flow adjustment policy.
  • Adjust the TCP flow generated when the application APP accesses the network specifically:
  • the TCP flow generated when the APP corresponding to the destination IP address that needs to be accelerated accesses the network is accelerated, and the destination IP address that needs to be speed-limited according to the need to be recorded in the TCP flow adjustment policy, and the purpose of speed-limiting that needs to be performed.
  • the rate of the TCP stream generated when the APP corresponding to the IP address accesses the network is limited.
  • this embodiment is a device embodiment corresponding to the first embodiment, and this embodiment can be implemented in cooperation with the first embodiment.
  • the technical details mentioned in the first embodiment are still valid in this embodiment, and are not repeated here in order to reduce repetition.
  • the technical details mentioned in this embodiment can also be applied in the first embodiment.
  • a logical unit may be a physical unit, a part of a physical unit, or multiple physical units.
  • a composite implementation of the unit in order to highlight the innovative part of the present application, this embodiment does not introduce units that are not closely related to solving the technical problem raised by the present application, but this does not mean that there are no other units in this embodiment.
  • the fifth embodiment of the present application relates to a TCP flow adjustment device.
  • the TCP flow adjustment device 700 includes a policy subscription module 701 , a policy reception module 702 and a TCP flow adjustment module 703 .
  • the policy subscription module 701 is configured to send a TCP flow adjustment policy subscription instruction to the NWDAF;
  • the policy receiving module 702 is configured to receive the TCP flow adjustment policy fed back by the NWDAF according to the TCP flow adjustment policy subscription instruction, the The TCP flow adjustment policy is determined by the NWDAF based on the SMF data fed back by the session management function SMF and the UPF data fed back by the UPF;
  • the TCP flow adjustment module 703 is configured to access the network for each application APP according to the TCP flow adjustment policy The resulting TCP stream is adjusted.
  • the TCP flow adjustment module 703 is specifically configured to read the destination IP address that needs to be accelerated and the destination IP address that needs to be speed-limited from the TCP flow adjustment policy; The destination IP address, to accelerate the TCP flow generated when the APP corresponding to the destination IP address that needs to be accelerated accesses the network; The TCP flow generated when the corresponding APP accesses the network is limited.
  • the method before reading the destination IP address that needs to be accelerated and the destination IP address that needs to be speed-limited from the TCP flow adjustment policy, the method further includes:
  • the TCP flow generated when the APP accesses the network is accelerated, and the destination IP address for speed-limiting according to the needs, the TCP flow generated when the APP corresponding to the destination IP address that needs to be speed-limited accesses the network is limited. quick steps;
  • this embodiment is a device embodiment corresponding to the second embodiment, and this embodiment can be implemented in cooperation with the second embodiment.
  • the technical details mentioned in the second embodiment are still valid in this embodiment, and are not repeated here in order to reduce repetition.
  • the related technical details mentioned in this embodiment can also be applied in the second embodiment.
  • a logical unit may be a physical unit, a part of a physical unit, or multiple physical units.
  • a composite implementation of the unit in order to highlight the innovative part of the present application, this embodiment does not introduce units that are not closely related to solving the technical problem raised by the present application, but this does not mean that there are no other units in this embodiment.
  • the memory 802 and the processor 801 are connected by a bus, and the bus may include any number of interconnected buses and bridges, and the bus connects one or more processors 801 and various circuits of the memory 802 together.
  • the bus may also connect together various other circuits, such as peripherals, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein.
  • the bus interface provides the interface between the bus and the transceiver.
  • a transceiver may be a single element or multiple elements, such as multiple receivers and transmitters, providing a means for communicating with various other devices over a transmission medium.
  • the data processed by the processor 801 is transmitted on the wireless medium through the antenna. In this embodiment, the antenna also receives the data and transmits the data to the processor 801 .
  • the processor 801 is responsible for managing the bus and general processing, and may also provide various functions including timing, peripheral interface, voltage regulation, power management, and other control functions.
  • the memory 802 may be used to store data used by the processor 801 when performing operations.
  • the seventh embodiment of the present application relates to a TCP acceleration apparatus, as shown in FIG. 9 , comprising: at least one processor 901 ; and a memory 902 communicatively connected with the at least one processor; wherein the memory 902 stores data that can be Instructions executed by the at least one processor 901, the instructions are executed by the at least one processor 901 to enable the at least one processor 901 to execute the adjustment method of the TCP stream applied to the TCP acceleration device.
  • the memory 902 and the processor 901 are connected by a bus, and the bus may include any number of interconnected buses and bridges, and the bus connects one or more processors 901 and various circuits of the memory 902 together.
  • the bus may also connect together various other circuits, such as peripherals, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein.
  • the bus interface provides the interface between the bus and the transceiver.
  • a transceiver may be a single element or multiple elements, such as multiple receivers and transmitters, providing a means for communicating with various other devices over a transmission medium.
  • the data processed by the processor 901 is transmitted on the wireless medium through the antenna, and further, the antenna also receives the data and transmits the data to the processor 901 .
  • Processor 901 is responsible for managing the bus and general processing, and may also provide various functions including timing, peripheral interface, voltage regulation, power management, and other control functions.
  • the memory 902 may be used to store data used by the processor 901 when performing operations.
  • the eighth embodiment of the present application relates to a computer-readable storage medium storing a computer program.
  • the computer program When the computer program is executed by the processor, it realizes the adjustment method of the TCP flow applied to the network data analysis function; or, the adjustment method of the TCP flow applied to the TCP acceleration device.
  • the aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes .

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer Security & Cryptography (AREA)
  • Computer Hardware Design (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本申请实施例涉及通信领域，公开了一种TCP流的调整方法和***。该方法应用于网络数据分析功能NWDAF，包括分别向会话管理功能SMF和用户面功能UPF发送日志订阅指令；接收所述SMF根据所述日志订阅指令反馈的SMF数据，以及所述UPF根据所述日志订阅指令反馈的UDF数据；对所述SMF数据和所述UDF数据进行分析，预估TCP流调整策略；将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略对各应用程序APP访问网络时产生的TCP流进行调整。

Description

TCP流的调整方法和***
相关申请的交叉引用
本申请基于申请号为“202011408143.8”、申请日为2020年12月04日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此以引入方式并入本申请。
技术领域
本申请实施例涉及通信领域，特别涉及一种TCP流的调整方法和***。
背景技术
5G通讯网络具有高速率(增强型移动宽带，eMBB)、大容量(大规模机器通信，mMTC)、低时延(高可靠低时延通信，uRLLC)的特性，因而随着5G通信网络的发展，大大提升了用户的网络访问体验。
但是，随着人们上网习惯由PC端转换至移动终端，再加上日益火热的视频、直播等，用户的移动网络带宽需求也增长迅速。为了提高运营商的竞争力，在这种情况下，如何利用有限的网络资源，尽可能提升移动终端的用户网络访问的体验显得尤为重要。
发明内容
本申请的实施例提供了一种TCP流的调整方法，应用于NWDAF(Network Data Analytics Function，网络数据分析功能)，所述方法包括：
SMF(Session management function，分别向会话管理功能)和UPF(User Plane Function，用户面功能)发送日志订阅指令；
接收所述SMF根据所述日志订阅指令反馈的SMF数据，以及所述UPF根据所述日志订阅指令反馈的UDF数据；
对所述SMF数据和所述UDF数据进行分析，预估TCP流调整策略；
将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略对各APP(Application，应用程序)访问网络时产生的TCP流进行调整。
本申请实施例还提供了一种TCP流的调整方法，应用于TCP加速装置，所述TCP加速装置分别与网络数据分析功能NWDAF、用户面功能UPF和外部网关通信连接，所述方法包括：
向所述NWDAF发送TCP流调整策略订阅指令；
接收所述NWDAF根据所述TCP流量调整策略订阅指令反馈的TCP流量调整策略，所述TCP流量调整策略由所述NWDAF基于会话管理功能SMF反馈的SMF数据和所述UPF反馈的UPF数据确定；
根据所述TCP流量调整策略，对各应用程序APP访问网络时产生的TCP流进行调整。
本申请实施例还提供了一种TCP流的调整装置，包括：
日志订阅模块，用于分别向会话管理功能SMF和用户面功能UPF发送日志订阅指令；
数据接收模块，用于接收所述SMF根据所述日志订阅指令反馈的SMF数据，以及所述UPF根据所述日志订阅指令反馈的UDF数据；
数据分析模块，用于对所述SMF数据和所述UDF数据进行分析，预估TCP流调整策略；
策略推送模块，用于将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略对各应用程序APP访问网络时产生的TCP流进行调整。
本申请实施例还提供了一种TCP流的调整装置，包括：
策略订阅模块，用于向所述NWDAF发送TCP流调整策略订阅指令；
策略接收模块，用于接收所述NWDAF根据所述TCP流量调整策略订阅指令反馈的TCP流量调整策略，所述TCP流量调整策略由所述NWDAF基于会话管理功能SMF反馈的SMF数据和所述UPF反馈的UPF数据确定；
TCP流调整模块，用于根据所述TCP流量调整策略，对各应用程序APP访问网络时产生的TCP流进行调整。
本申请实施例还提供了一种网络数据分析功能，包括：
至少一个处理器；以及，
与所述至少一个处理器通信连接的存储器；其中，
所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行如上所述应用于NWDAF的TCP流的调整方法。
本申请实施例还提供了一种TCP加速装置，包括：
至少一个处理器；以及，
与所述至少一个处理器通信连接的存储器；其中，
所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行如上所述应用于TCP加速装置的TCP流的调整方法。
本申请实施例还提供了一种计算机可读存储介质，存储有计算机程序，所述计算机程序被处理器执行时实现如上所述应用于NWDAF的TCP流的调整方法，或者如上所述应用于TCP加速装置的TCP流的调整方法。
本申请实施例还提供了一种TCP流的调整***，包括：
如上所述的TCP加速装置；以及，
与所述TCP加速装置通信连接的如上所述的网络数据分析功能NWDAF、用户面功能UPF和外部网关；以及，
与所述NWDAF通信连接的会话管理功能SMF；其中，
所述NWDAF还与所述UPF通信连接；
所述NWDAF用于执行如上所述应用于NWDAF的TCP流的调整方法；
所述TCP加速装置用于执行如上所述应用于TCP加速装置的TCP流的调整方法。
附图说明
一个或多个实施例通过与之对应的附图中的图片进行示例性说明，这些示例性说明并不构成对实施例的限定。
图1是本申请第一实施例提供的TCP流的调整***的结构示意图；
图2是本申请第一实施例提供的TCP流的调整***在5G网络中的架构示意图；
图3是本申请第二实施例提供的TCP流的调整方法的流程图；
图4是本申请第二实施例提供的TCP流的调整方法涉及的各功能实体的交互示意图；
图5是本申请第三实施例提供的TCP流的调整方法的流程图；
图6是本申请第四实施例提供的TCP流的调整装置的结构示意图；
图7是本申请第五实施例提供的TCP流的调整装置的结构示意图；
图8是本申请第六实施例提供的网络数据分析功能的结构示意图；
图9是本申请第七实施例提供的TCP加速装置的结构示意图。
具体实施方式
本申请实施例提供一种TCP(Transmission Control Protocol，传输控制协议)流的调整方法和***，NWDAF(Network Data Analytics Function)是一个数据感知分析网元，以网络数据为基础对网络进行自动感知和分析，并参与到网络规划、建设、运维、网优、运营全生命周期中，使得网络易于维护和控制，提高网络资源使用效率，提升用户业务体验。基于此，本申请提出的TCP流的调整方法和***，通过利用NWDAF的智能分析统计加上趋势预测，对SMF(Session management function)提供的SMF数据和UPF(User Plane Function)提供的UPF数据进行分析，进而预估出对各APP访问网络时产生的TCP流的调整策略，如对某类APP的TCP流进行加速，某类则进行限速，最终将预估出的TCP流调整策略推送至预先接入UPF和外部网关之间的TCP加速装置，从而可以使TCP加速装置能够对来自UPF的TCP流进行合理的调整，进而在有限的网络资源下，尽可能的提升了移动终端的用户网络访问的体验。
为使本申请实施例的目的、技术方案和优点更加清楚，下面将结合附图对本申请的各实施例进行详细的阐述。然而，本领域的普通技术人员可以理解，在本申请各实施例中，为了使读者更好地理解本申请而提出了许多技术细节。但是，即使没有这些技术细节和基于以下各实施例的种种变化和修改，也可以实现本申请所要求保护的技术方案。以下各个实施例的划分是为了描述方便，不应对本申请的具体实现方式构成任何限定，各个实施例在不矛盾的前提下可以相互结合相互引用。
本申请的第一实施例涉及一种TCP流的调整***，如图1所示。TCP流的调整***，包括：TCP加速装置10、网络数据分析功能20、会话管理功能30、用户面功能40和外部网关50。
其中，TCP加速装置10分别与网络数据分析功能20、用户面功能40和外部网关50通信连接，网络数据分析功能20又与会话管理功能30和用户面功能40通信连接。
可理解的是，本实施例中所说的网络数据分析功能20，即通信领域中所说的NWDAF(Network Data Analytics Function)。它是第五代移动通信技术(5th-Generation，5G)***提出的一种功能实体，可以看作一个数据感知分析网元，主要以网络数据为基础对网络进行 自动感知和分析，并参与到网络规划、建设、运维、网优、运营全生命周期中，使得网络易于维护和控制，提高网络资源使用效率，提升用户业务体验。
同样，会话管理功能30，也是5G基于服务架构的一个功能实体，即通信领域中所说的SMF(Session management function)。具体的，在实际应用中，SMF主要负责与分类的数据面交互，创建、更新和删除协议数据单元(Protocol Data Unit，PDU)会话，并管理与用户面功能40的会话环境(session context)，即在实际应用中，会话管理功能30和用户面功能40之间也是存在通信连接的。
同样，用户面功能40，也是5G基于服务架构中的一个重要功能实体，即通信领域中所说的UPF(User Plane Function)。
而TCP加速装置10，则是本实施例新加入5G基于服务架构中的一个功能实体，主要用于根据网络数据分功能20提供的TCP流调整策略对TCP流进行加速或限速的调整。
为了更好的理解本实施例提供的TCP流的调整策略对TCP流的调整，以下结合图2给出的5G网络中的架构图进行说明。
为了便于说明，以下先对图2所示的架构中的网元功能实体进行简要说明：
如图2所示，NSSF(Network Slice Selection Function，网络切片选择功能)、AUSF(Authentication Server Function，鉴权服务功能)、NEF(Network Exposure Function，网络业务呈现功能)、UDM(Unified Data Management，统一数据管理功能)、NRF(NF Repository Function，网元数据仓库功能)、PCE(Policy Control function，策略控制功能)、AF(Application Function，应用层功能)这些网元功能实体通过网络总线与AMF(Access and Mobility Management Function，接入及移动性管理功能)、SMF(Session management function，会话管理功能)和NWDAF(Network Data Analytics Function，网络数据分析功能)这些网元功能实体连通，AMF则通过N1接口(UE和AMF间的信令面接口，是逻辑接口)与UE(User Equipment，用户终端设备)连通，通过N2接口((R)AN和AMF间的信令面接口，类似于4G中的s1-mme接口)与(R)AN(Radio Access Network，接入网络)连通，UE通过有线或无线的方式接入(R)AN，(R)AN通过N3接口((R)AN与UPF间的接口，主要用于传递5G(R)AN与UPF间的上下行用户面数据)与UPF连通，UPF通过N4接口(SMF和UPF间的接口，用于传输SMF和UPF间的控制面信息)与SMF连通，TCPO(TCP加速装置)通过N6接口(UPF与DN的接口，用于传递UPF与DN之间的上下行用户数据流，基于IP和路由协议与DN网络通信)分别与UPF和DN(Data Network，数据网络，比如运营商业务，互联网接入或第三方业务等)连通。
需要说明的是，上述所说的DN即为本实施例中所说的外部网关，在实际应用中，外部网关除了可以是DN之外，还可以是FW(firewall，防火墙)，或者核心网中的其他承载了网关作用的网络设备。
此外，应当理解的是，以上给出的仅为5G网络中目前的一种具体架构，对本申请的技术方案并不构成限定，在实际应用中，架构中的具体网元功能实体可以根据业务需要有所变动，只要确保架构中含有本实施例中所说的TCP加速装置10、网络数据分析功能20、会话管理功能30、用户面功能40和外部网关50，并且保证这5个网元功能实体之间，按照本实施例中所说的连接关系进行通信连通即可。
基于图2所示的结构，在实现对TCP流的调整时，具体为：
首先，由TCPO(即图1中的TCP加速装置10)向NWDAF(即图1中的网络数据分析功能20)发送TCP流调整策略订阅指令。
接着，NWDAF在接收到TCPO发送来的TCP流调整策略订阅指令后，生成日志订阅指令，并将生成的日志订阅指令，分别发送至SMF(即图1中的会话管理功能30)和UPF(即图1中的用户面功能40)，从而使得SMF能够在监测到有创建，或者更新，或者释放PDU会话时，能够将产生的数据(为了便于描述，以下称为SMF数据)推送至NWDAF；使得UF能够在监测到流释放操作时，将产生的数据(为了便于描述，以下称为UPF数据)推送至NWDAF。
接着，NWDAF基于其能够对网络数据进行分析和预测的能力，对SMF数据和UDF数据进行分析，进而预估出适合当前网络TCP流调整策略，并将得到TCP流调整策略推送至TCPO。
最后，TCPO根据接收到的TCP流调整策略，对UPF释放出的各APP(Application，应用程序)访问网络时产生的TCP流进行调整，如对高价值APP对应的TCP流进行加速，然后发往外部网关DN，对低价值APP对应的TCP流进行限速，然后发往外部网关DN。
由此，本实施例提供的TCP流的调整***，通过利用NWDAF的智能分析统计加上趋势预测，对SMF提供的SMF数据和UPF提供的UPF数据进行分析，进而预估出对各APP访问网络时产生的TCP流的调整策略，如对某类APP的TCP流进行加速，某类则进行限速，最终将预估出的TCP流调整策略推送至预先接入UPF和外部网关之间的TCP加速装置，从而可以使TCP加速装置能够对来自UPF的TCP流进行合理的调整，进而在有限的网络资源下，尽可能的提升了移动终端的用户网络访问的体验。
本申请的第二实施例涉及一种TCP流的调整方法，应用于网络数据分析功能NWDAF。
关于NWDAF，与实现TCP流的调整方法时所涉及的其他网元功能实体，以及与这些网元功能实体之间的连接关系，详见本申请第一实施例提供的TCP流的调整***所对应的图1和图2。
下面对本实施例的TCP流的调整方法的实现细节进行说明，以下内容仅为方便理解而提供的实现细节，并非实施本方案的必须。
本实施例的具体流程如图3所示，具体包括以下步骤：
步骤301，分别向会话管理功能SMF和用户面功能UPF发送日志订阅指令。
步骤302，接收所述SMF根据所述日志订阅指令反馈的SMF数据，以及所述UPF根据所述日志订阅指令反馈的UDF数据。
步骤303，对所述SMF数据和所述UDF数据进行分析，预估TCP流调整策略。
步骤304，将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略对各应用程序APP访问网络时产生的TCP流进行调整。
为了更好的理解步骤301至步骤304中的操作，以下结合图4进行具体说明：
(1)TCPO(TCP加速装置)向NWDAF发送TCP流调整策略订阅指令。
即，在实际应用中，触发NWDAF执行步骤301至步骤304操作的前提是，NWDAF接收到了TCPO(TCP加速装置)发来的TCP流调整策略订阅指令。
(2)NWDAF在接收到TCPO发送来的TCP流调整策略订阅指令后，便会生成日志订阅指令，并将生成的日志订阅指令分别发送至SMF和UPF。
关于上述所说的日志订阅指令，在实际应用中，具体是指的订阅呼叫历史记录日志(call history record日志，CHR日志)的指令，即后续接收到的SMF根据该日志订阅指令反馈的SMF数据和UPF根据该日志订阅指令反馈的UPF数据时从CHR日志中提取出的。
通过第一实施例中对SMF的介绍可知，SMF主要负责与分类的数据面交互，创建、更新和删除PDU会话。故而，在有创建，或者更新，或者删除等交互操作产生的PDU会话时，SMF便会进行控制面创建信息采集的操作，并将采集到的数据(以下称为SMF数据)推送至NWDAF。
相应地，由于UPF主要负责将UE，本实施例中主要指移动用户设备中的APP访问网络时产生的流，本实施例中主要指TCP流发往外部网关，即进行流释放。故而，在有流释放时，UPF便会进行流信息采集，并将采集到的数据(以下称为UPF数据)推送至NWDAF。
(3)NWDAF在接收到SMF数据和UPF数据之后，便会基于自身的分析和预测能力，对所述SMF数据和所述UDF数据进行分析，进而预估得到TCP流调整策略。
在本实施例中，关于NWDAF对所述SMF数据和所述UDF数据进行分析，进而预估得到TCP流调整策略的操作，在实际应用中，具体为：
首先，将所述SMF数据和所述UDF数据，按照预设字段信息进行合并，进而得到待处理数据。
在本实施例中，具体是将SEID(近场通信NFC的序列号)和N4接口作为预设字段信息，将SMF数据和UDF数据进行合并，即将N4接口提供的，SEID相同的数据进行合并。
基于这种方式，合并后的待处理数据，主要包括：国际移动用户识别码(International Mobile Subscriber Identity，IMSI)、私网地址、私网端口、协议类型、目的地址、目的端口，包数，时长，已知的APP类型，统一资源定位***(uniform resource locator，URL)等。
然后，基于预设的趋势预测机器学习算法，以预设时间粒度和预设区域为维度，对所述待处理数据中不同时间维度、不同使用区域内的数据进行分析，得到所述TCP流调整策略。
在本实施例中，预估TCP流调整策略，是为了能够实现快速识别占用流量大，但价值低的APP，以及高价值，如热门APP。
故而，在基于上述方式，预估TCP流调整策略时，具体是以预设时间粒度和预设区域为维度，统计所述待处理数据中不同时间维度、不同使用区域内，所述各用户端的源IP地址、所述各用户端中所述各APP对应的目的IP地址，以及所述各APP的历史流量使用信息；接着，通过遍历所述各APP，建立所述源IP地址与遍历到的APP对应的目的IP地址、历史流量使用信息之间的映射关系，得到所述各APP的历史流量使用统计表；最终，基于预设的趋势预测机器学习算法，对所述历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，得到所述TCP流调整策略。
关于上述所说的预设时间粒度，可以是以分钟、小时、天为单位，即每一分钟，或者一小时，或者一天统计一次；所说预设区域，可以是以小区为单位。
此外，需要说明的是，由于在实际应用中，用于对TCP流进行调整的TCP加速装置，只能识别目的IP，通过目的IP对相应的TCP流进行加速或限速调整。故而，上述所说的TCP流调整策略中，具体记录的是需要进行加速的目的IP地址和需要进行限速的目的IP地址。
关于上述所说的需要进行限速的目的IP地址，具体是由NWDAF基于预设的趋势预测机器学习算法，通过对历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，进而识别出的占用流量大的低价值APP对应的目的IP地址。
需要说明的是，本实施例中所说的占用流量大的低价值APP，具体是由NWDAF基于待历史流量使用统计表中记录的数据，通过分析统计得出的，出现网络拥塞现象的前N个小区的历史流量使用情况，以及出现拥塞时，频繁出现的UE的IMSI和占用流量大但是使用占比低的APP。
比如说，通过统计发现，在当前统计时间内，出现拥堵现象，并属于前N个小区中的某一个小区内有1000个用户接入网络，但是其中只有10个用户，即10个不同IMSI的UE在使用P2P类型的APP访问网络，但却占用了分配给该小区的大量流量。此时，为了不影响其他用户群体(其他990)的上网体验，便会认为这0个不同IMSI的UE在使用P2P类型的APP为占用流量大的低价值APP，即需要进行限速的。
相应地，上述所说的需要进行加速的目的IP地址，具体是由NWDAF基于预设的趋势预测机器学习算法，通过对历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，进而识别出的高价值APP对应的目的IP地址。
关于上述所说的高价值APP，在实际应用中，可以根据用户热度确定，比如用户访问量高的APP，可以认为是热门APP。对于这类APP便可以认为是高价值APP。
应当理解的是，上述示例仅是为了更好的理解本实施例的技术方案而列举的示例，不作为对本实施例的唯一限制。
本实施例中，考虑到季节以及节假日等因素对预测结果的影响，即对于某一类APP的高频访问，可能仅仅出现在特定季节或节假日，如冬季冷，大部分用户群体喜欢宅在家追剧，这一阶段视频类型的APP就可能成为热门APP，还比如对于电商在某一特定时间进行的线上促销活动，可能在这段时间的固定时间点，直播类型的软件由于会进行线上促销，则此类APP就可能成为热门APP。故而，为了使得预估的TCP流调整策略更符合实际情况，NWDAF在预估TCP流调整策略时，可以先获取所述历史流量使用统计表中记录的每一条存在映射关系的数据中所说历史流量使用信息的产生时间，然后对于所述历史流量使用统计表中记录的每一条存在映射关系的数据，基于所述趋势预测机器学习算法，结合所述产生时间进行分析，得到所述TCP流调整策略。
由此，通过在预估TCP流调整策略时，引入各APP的历史流量使用信息的产生时间，从而可以根据产生时间确定当前的季节，以及是否为节假日、某些特定活动日等，进而在基于趋势预测机器学习算法对历史流量使用统计表中的数据进行分析时，能够考虑到季节、节假日等因素对趋势造成的影响，使得预估出的TCP流调整策略更加合理。
此外，值得一提的是，关于上述所说的趋势预测机器学***均模型(Autoregressive Integrated Moving Average Model，ARIMA)，prophet(时间序列预测库)模型，长短时间记忆模型(Long Short-Term Memory，LSTM)等，此处不再一一列举，本实施例对此也不做限制。
为了便于理解，以下以LSTM模型为例，对预估TCP流调整策略的操作进行说明：
具体的说，在实际应用中，为了能够给快速、准确的预估出适合实际情况的TCP流调整策略，可以预先构建有一个LSTM和一个全连接层的网络模型，然后通过对构建的网络模型 进行迭代训练，进而得到一个TCP流调整策略预估模型，这样在每次获取到新的SMF数据和UDP数据时，便可以直接将二者合并后输入到TCP流调整策略预估模型，进而将TCP流调整策略预估模型输出的结果作为预估出的TCP流调整策略。
关于TCP流调整策略预估模型的训练，具体如下：
首先，将预先获取的样本数据，即由历史时刻产生的SMF数据和UDP数据合并而成的数据，按照7：3的比例划分为训练集和测试集。
然后，对固定时间的时间段的训练集中的流量值进行归一化处理，从而提升梯度项下降速度。
接着，将归一化处理后的流量值按序输入构建的网络模型进行迭代训练，例如：输入的流量值分别为t-4，t-3，t-2，t-1时刻的流量值，基于这几个时刻的流量值，预测t时刻的流量值，再将预测出的t时刻的流量值加入训练集，预测t+1时刻的流量值，总体过程如下：
输入：t-4，t-3，t-2，t-1；预测：t
输入：t-3，t-2，t-1，t；预测：t+1
输入：t-2，t-1，t-0，t+1；预测：t+2
整个过程不断迭代，直到达到需要预测的粒度即可。
接着，将上述时间序列输入网络模型后输出的预测值，与实际的真实值进行比较，采用均方误差MSE损失函数，用ADAM优化算法来优化损失函数，通过优化损失函数，不断的进行梯度下降训练，随着损失函数的收敛，便可以得到合适的网络模型。
最后，同样经过归一化处理后的测试集中，不同时刻对应的流量值输入经过迭代训练后的网络模型，然后将输出的预测值，与实际的真实值进行比较，当测试结果的性能指标达到预期时，如测试值基本与真实值相同，或者误差在某一范围内，便可以将当前的网络模型确定为TCP流调整策略预估模型，即后续可以使用该模型来预测未来某一时刻，该小区的流量趋势，进入制定出适合该小区后续的TCP流调整策略。
比如，一旦预测出该小区背后的流量趋势呈现出明显的上升趋势，如上升趋势的增长率超过预设值，例如30min内流量上升30％，则将从历史数据统计到的导致小区拥塞的IMSI对应的私网IP及其低价值APP访问的目的IP确定为需要进行限速的目的IP，并将这一TCP流调整策略推送给TCP加速装置，由TCP加速装置根据需要进行限速的目的IP，对这一目的IP对应的APP访问网络时产生的TCP流进行限速，反之则由TCP加速装置根据需要进行加速的目的IP，对这一目的IP对应的APP访问网络时产生的TCP流进行加速。
也就是说，图4中，TCPO在接收到NWDAF推送来的TCP流调整策略后，会根据所述TCP流调整策略中记录的需要进行加速的目的IP地址，识别UPF发送来的UE互联网访问流中的目的IP地址，进而将需要进行加速的目的IP地址，即高价值APP对应的TCP流进行加速，然后将加速后的TCP流发往外部网关；根据所述TCP流调整策略中记录的需要进行限速的目的IP地址，识别UPF发送来的UE互联网访问流中的目的IP地址，进而将需要进行限速的目的IP地址，即低价值APP对应的TCP流进行限速，然后将限速后的TCP流发往外部网关。
通过上述描述不难发现，本实施例提供的TCP流的调整方法，通过利用NWDAF的智能分析统计加上趋势预测，对SMF提供的SMF数据和UPF提供的UPF数据进行分析，进而预估出对各APP访问网络时产生的TCP流的调整策略，如对某类APP的TCP流进行加速， 某类则进行限速，最终将预估出的TCP流调整策略推送至预先接入UPF和外部网关之间的TCP加速装置，从而可以使TCP加速装置能够对来自UPF的TCP流进行合理的调整，进而在有限的网络资源下，尽可能的提升了移动终端的用户网络访问的体验。
也就是说，基于本实施例提供的TCP流的调整方法，可以实现在移动网络忙碌的情况下，通过调整数据流包的策略，如对低价值APP的TCP流进行限速，对高价值APP的TCP流进行加速的调整策略，对各APP访问网络时产生的TCP流进行合理的调整，而不是对全部的TCP流进行统一的调整，如均进行限速或加速，使得有限的移动网络资源能够得到合理的利用，从而提升用户体验，增加运营商用户粘性，从而增强运营商竞争力。
本申请的第三实施例涉及一种TCP流的调整方法，应用于TCP加速装置。
关于TCP加速装置，与实现TCP流的调整方法时所涉及的其他网元功能实体，以及与这些网元功能实体之间的连接关系，详见本申请第一实施例提供的TCP流的调整***所对应的图1和图2。
下面对本实施例的TCP流的调整方法的实现细节进行说明，以下内容仅为方便理解而提供的实现细节，并非实施本方案的必须。
如图5所示，第二实施例涉及的TCP流的调整方法，包括如下步骤：
步骤501，向所述NWDAF发送TCP流调整策略订阅指令。
步骤502，接收所述NWDAF根据所述TCP流量调整策略订阅指令反馈的TCP流量调整策略，所述TCP流量调整策略由所述NWDAF基于会话管理功能SMF反馈的SMF数据和所述UPF反馈的UPF数据确定。
步骤503，根据所述TCP流量调整策略，对各应用程序APP访问网络时产生的TCP流进行调整。
本实施例中，由于通过NWDAF预估出的TCP流调整策略中记录了需要进行加速的目的IP地址，以及需要进行限速的目的IP地址。
故而，在根据TCP流调整策略，对各APP访问网络时产生的TCP流进行调整时，具体是通过从TCP流调整策略中读取需要进行加速的目的IP地址和需要进行限速的目的IP地址，然后基于需要进行加速的目的IP地址，对需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速。
相应地，基于需要进行限速的目的IP地址，对需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速。
本实施例中，由于在实际应用中，可能存在NWDAF基于SMF数据和UPF数据进行分析，预估出的TCP流调整策略中仅包括加速或限速一种对TCP流的调整信息。
故而，为了使得TCP加速装置能够更好的适应于实际使用场景，在TCP加速装置从TCP流调整策略中读取需要进行加速的目的IP地址和需要进行限速的目的IP地址之前，可以先判断所述TCP流调整策略中是否携带了所述需要进行加速的目的IP地址和所述需要进行限速的目的IP地址，然后根据判断结果触发后续的调整操作。
本实施例中，若通过判断，确定TCP流调出策略中携带了所述需要进行加速的目的IP地址和所述需要进行限速的目的IP地址，则执行所述根据所述需要进行加速的目的IP地址，对所述需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速，以及所 述根据所述需要进行限速的目的IP地址，对所述需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速的步骤；
若通过判断，确定TCP流调出策略中仅携带了所述需要进行加速的目的IP地址，则仅执行所述根据所述需要进行加速的目的IP地址，对所述需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速的步骤；
若通过判断，确定TCP流调出策略中仅携带了所述需要进行限速的目的IP地址，则仅执行所述根据所述需要进行限速的目的IP地址，对所述需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速的步骤。
不难发现，本实施例中的步骤501至步骤503，是与与第二实施例中的步骤301至步骤304存在关联关系，即在实现TCP流的调整时，整体方案的实现需要涉及步骤301至步骤304，以及步骤501至步骤503的全部操作。关于本实施例中涉及的步骤501至步骤503的具体实现，已经在第二实施例中结合图4的介绍中说明，在此就不再赘述。
由此，本实施例提供的TCP流的调整方法，通过利用NWDAF的智能分析统计加上趋势预测，对SMF提供的SMF数据和UPF提供的UPF数据进行分析，进而预估出对各APP访问网络时产生的TCP流的调整策略，如对某类APP的TCP流进行加速，某类则进行限速，最终将预估出的TCP流调整策略推送至预先接入UPF和外部网关之间的TCP加速装置，从而可以使TCP加速装置能够对来自UPF的TCP流进行合理的调整，进而在有限的网络资源下，尽可能的提升了移动终端的用户网络访问的体验。
此外，应当理解的是，上面各种方法的步骤划分，只是为了描述清楚，实现时可以合并为一个步骤或者对某些步骤进行拆分，分解为多个步骤，只要包括相同的逻辑关系，都在本专利的保护范围内；对算法中或者流程中添加无关紧要的修改或者引入无关紧要的设计，但不改变其算法和流程的核心设计都在该专利的保护范围内。
本申请的第四实施例涉及一种TCP流的调整装置，如图6所示，TCP流的调整装置600包括：日志订阅模块601、数据接收模块602、数据分析模块603和策略推送模块604。
其中，日志订阅模块601，用于分别向会话管理功能SMF和用户面功能UPF发送日志订阅指令；数据接收模块602，用于接收所述SMF根据所述日志订阅指令反馈的SMF数据，以及所述UPF根据所述日志订阅指令反馈的UDF数据；数据分析模块603，用于对所述SMF数据和所述UDF数据进行分析，预估TCP流调整策略；策略推送模块604，用于将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略对各应用程序APP访问网络时产生的TCP流进行调整。
此外，在另一个例子中，数据分析模块603具体用于将所述SMF数据和所述UDF数据，按照预设字段信息进行合并，得到待处理数据；基于预设的趋势预测机器学习算法，以预设时间粒度和预设区域为维度，对所述待处理数据中不同时间维度、不同使用区域内的数据进行分析，得到所述TCP流调整策略。
此外，在另一个例子中，所述基于预设的趋势预测机器学习算法，以预设时间粒度和预设区域为维度，对所述待处理数据中不同时间维度、不同使用区域内的数据进行分析，得到所述TCP流调整策略，具体为：
以预设时间粒度和预设区域为维度，统计所述待处理数据中不同时间维度、不同使用区域内，所述各用户端的源IP地址、所述各用户端中所述各APP对应的目的IP地址，以及所 述各APP的历史流量使用信息；
遍历所述各APP，建立所述源IP地址与遍历到的APP对应的目的IP地址、历史流量使用信息之间的映射关系，得到所述各APP的历史流量使用统计表；
基于预设的趋势预测机器学习算法，对所述历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，得到所述TCP流调整策略。
此外，在另一个例子中，所述基于预设的趋势预测机器学习算法，对所述历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，得到所述TCP流调整策略，具体为：
获取所述历史流量使用统计表中记录的每一条存在映射关系的数据中所说历史流量使用信息的产生时间；
对于所述历史流量使用统计表中记录的每一条存在映射关系的数据，基于所述趋势预测机器学习算法，结合所述产生时间进行分析，得到所述TCP流调整策略。
此外，在另一个例子中，所述基于预设的趋势预测机器学习算法，对所述历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，得到所述TCP流调整策略，具体为：
基于预设的趋势预测机器学习算法，对所述历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，识别出占用流量大的低价值APP和高价值APP；
获取所述低价值APP对应的目的IP地址和所述高价值APP对应的目的IP地址；
将所述低价值APP对应的目的IP地址作为需要进行限速的目的IP地址，将所述高价值APP对应的目的IP地址作为需要进行加速的目的IP地址，得到所述TCP流调整策略。
此外，在另一个例子中，所述将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略对各应用程序APP访问网络时产生的TCP流进行调整，具体为：
将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略中记录的需要进行加速的目的IP地址，对所述需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速，根据所述TCP流调整策略中记录的需要进行限速的目的IP地址，对所述需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速。
不难发现，本实施例为与第一实施例相对应的装置实施例，本实施例可与第一实施例互相配合实施。第一实施例中提到的技术细节在本实施例中依然有效，为了减少重复，这里不再赘述。相应地，本实施例中提到的技术细节也可应用在第一实施例中。
值得一提的是，本实施例中所涉及到的各模块均为逻辑模块，在实际应用中，一个逻辑单元可以是一个物理单元，也可以是一个物理单元的一部分，还可以以多个物理单元的组合实现。此外，为了突出本申请的创新部分，本实施例中并没有将与解决本申请所提出的技术问题关系不太密切的单元引入，但这并不表明本实施例中不存在其它的单元。
本申请的第五实施例涉及一种TCP流的调整装置，如图7所示，TCP流的调整装置700包括：策略订阅模块701、策略接收模块702和TCP流调整模块703。
其中，策略订阅模块701，用于向所述NWDAF发送TCP流调整策略订阅指令；策略接收模块702，用于接收所述NWDAF根据所述TCP流量调整策略订阅指令反馈的TCP流量调整策略，所述TCP流量调整策略由所述NWDAF基于会话管理功能SMF反馈的SMF数据和所述UPF反馈的UPF数据确定；TCP流调整模块703，用于根据所述TCP流量调整策略， 对各应用程序APP访问网络时产生的TCP流进行调整。
此外，在另一个例子中，TCP流调整模块703具体用于从所述TCP流调整策略中读取需要进行加速的目的IP地址和需要进行限速的目的IP地址；根据所述需要进行加速的目的IP地址，对所述需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速；根据所述需要进行限速的目的IP地址，对所述需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速。
此外，在另一个例子中，在所述从所述TCP流调整策略中读取需要进行加速的目的IP地址和需要进行限速的目的IP地址之前，还包括：
判断所述TCP流调整策略中是否携带了所述需要进行加速的目的IP地址和所述需要进行限速的目的IP地址；
若携带了所述需要进行加速的目的IP地址和所述需要进行限速的目的IP地址，则执行所述根据所述需要进行加速的目的IP地址，对所述需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速，以及所述根据所述需要进行限速的目的IP地址，对所述需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速的步骤；
若仅携带了所述需要进行加速的目的IP地址，则仅执行所述根据所述需要进行加速的目的IP地址，对所述需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速的步骤；
若仅携带了所述需要进行限速的目的IP地址，则仅执行所述根据所述需要进行限速的目的IP地址，对所述需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速的步骤。
不难发现，本实施例为与第二实施例相对应的装置实施例，本实施例可与第二实施例互相配合实施。第二实施例中提到的技术细节在本实施例中依然有效，为了减少重复，这里不再赘述。相应地，本实施例中提到的相关技术细节也可应用在第二实施例中。
值得一提的是，本实施例中所涉及到的各模块均为逻辑模块，在实际应用中，一个逻辑单元可以是一个物理单元，也可以是一个物理单元的一部分，还可以以多个物理单元的组合实现。此外，为了突出本申请的创新部分，本实施例中并没有将与解决本申请所提出的技术问题关系不太密切的单元引入，但这并不表明本实施例中不存在其它的单元。
本申请的第六实施例涉及一种网络数据分析功能，如图8所示，包括：包括至少一个处理器801；以及，与至少一个处理器通信连接的存储器802；其中，存储器802存储有可被至少一个处理器801执行的指令，指令被至少一个处理器801执行，以使至少一个处理器801能够执行上述方应用于网络数据分析功能的TCP流的调整方法。
其中，存储器802和处理器801采用总线方式连接，总线可以包括任意数量的互联的总线和桥，总线将一个或多个处理器801和存储器802的各种电路连接在一起。总线还可以将诸如***设备、稳压器和功率管理电路等之类的各种其他电路连接在一起，这些都是本领域所公知的，因此，本文不再对其进行进一步描述。总线接口在总线和收发机之间提供接口。收发机可以是一个元件，也可以是多个元件，比如多个接收器和发送器，提供用于在传输介质上与各种其他装置通信的单元。经处理器801处理的数据通过天线在无线介质上进行传输，本实施例中，天线还接收数据并将数据传送给处理器801。
处理器801负责管理总线和通常的处理，还可以提供各种功能，包括定时，***接口， 电压调节、电源管理以及其他控制功能。而存储器802可以被用于存储处理器801在执行操作时所使用的数据。
本申请的第七实施例涉及一种TCP加速装置，如图9所示，包括：包括至少一个处理器901；以及，与至少一个处理器通信连接的存储器902；其中，存储器902存储有可被至少一个处理器901执行的指令，指令被至少一个处理器901执行，以使至少一个处理器901能够执行应用于TCP加速装置的TCP流的调整方法。
其中，存储器902和处理器901采用总线方式连接，总线可以包括任意数量的互联的总线和桥，总线将一个或多个处理器901和存储器902的各种电路连接在一起。总线还可以将诸如***设备、稳压器和功率管理电路等之类的各种其他电路连接在一起，这些都是本领域所公知的，因此，本文不再对其进行进一步描述。总线接口在总线和收发机之间提供接口。收发机可以是一个元件，也可以是多个元件，比如多个接收器和发送器，提供用于在传输介质上与各种其他装置通信的单元。经处理器901处理的数据通过天线在无线介质上进行传输，进一步，天线还接收数据并将数据传送给处理器901。
处理器901负责管理总线和通常的处理，还可以提供各种功能，包括定时，***接口，电压调节、电源管理以及其他控制功能。而存储器902可以被用于存储处理器901在执行操作时所使用的数据。
本申请的第八实施例涉及一种计算机可读存储介质，存储有计算机程序。计算机程序被处理器执行时实现应用于网络数据分析功能的TCP流的调整方法；或者，应用于TCP加速装置的TCP流的调整方法。
即，本领域技术人员可以理解，实现上述实施例方法中的全部或部分步骤是可以通过程序来指令相关的硬件来完成，该程序存储在一个存储介质中，包括若干指令用以使得一个设备(可以是单片机，芯片等)或处理器(processor)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
本领域的普通技术人员可以理解，上述各实施例是实现本申请的具体实施例，而在实际应用中，可以在形式上和细节上对其作各种改变，而不偏离本申请的精神和范围。

Claims (15)

  1. 一种TCP流的调整方法，应用于网络数据分析功能NWDAF，所述方法包括：
    分别向会话管理功能SMF和用户面功能UPF发送日志订阅指令；
    接收所述SMF根据所述日志订阅指令反馈的SMF数据，以及所述UPF根据所述日志订阅指令反馈的UDF数据；
    对所述SMF数据和所述UDF数据进行分析，预估TCP流调整策略；
    将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略对各应用程序APP访问网络时产生的TCP流进行调整。
  2. 如权利要求1所述的TCP流的调整方法，其中，所述对所述SMF数据和所述UDF数据进行分析，预估TCP流调整策略，包括：
    将所述SMF数据和所述UDF数据，按照预设字段信息进行合并，得到待处理数据；
    基于预设的趋势预测机器学习算法，以预设时间粒度和预设区域为维度，对所述待处理数据中不同时间维度、不同使用区域内的数据进行分析，得到所述TCP流调整策略。
  3. 如权利要求2所述的TCP流的调整方法，其中，所述基于预设的趋势预测机器学习算法，以预设时间粒度和预设区域为维度，对所述待处理数据中不同时间维度、不同使用区域内的数据进行分析，得到所述TCP流调整策略，包括：
    以预设时间粒度和预设区域为维度，统计所述待处理数据中不同时间维度、不同使用区域内，所述各用户端的源IP地址、所述各用户端中所述各APP对应的目的IP地址，以及所述各APP的历史流量使用信息；
    遍历所述各APP，建立所述源IP地址与遍历到的APP对应的目的IP地址、历史流量使用信息之间的映射关系，得到所述各APP的历史流量使用统计表；
    基于预设的趋势预测机器学习算法，对所述历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，得到所述TCP流调整策略。
  4. 如权利要求3所述的TCP流的调整方法，其中，所述基于预设的趋势预测机器学习算法，对所述历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，得到所述TCP流调整策略，包括：
    获取所述历史流量使用统计表中记录的每一条存在映射关系的数据中所说历史流量使用信息的产生时间；
    对于所述历史流量使用统计表中记录的每一条存在映射关系的数据，基于所述趋势预测机器学习算法，结合所述产生时间进行分析，得到所述TCP流调整策略。
  5. 如权利要求3或4所述的TCP流的调整方法，其中，所述基于预设的趋势预测机器学习算法，对所述历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，得到所述TCP流调整策略，包括：
    基于预设的趋势预测机器学习算法，对所述历史流量使用统计表中记录的每一条存在映射关系的数据进行分析，识别出占用流量大的低价值APP和高价值APP；
    获取所述低价值APP对应的目的IP地址和所述高价值APP对应的目的IP地址；
    将所述低价值APP对应的目的IP地址作为需要进行限速的目的IP地址，将所述高价值 APP对应的目的IP地址作为需要进行加速的目的IP地址，得到所述TCP流调整策略。
  6. 如权利要求5所述的TCP流的调整方法，其中，所述将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略对各应用程序APP访问网络时产生的TCP流进行调整，包括：
    将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略中记录的需要进行加速的目的IP地址，对所述需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速，根据所述TCP流调整策略中记录的需要进行限速的目的IP地址，对所述需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速。
  7. 一种TCP流的调整方法，应用于TCP加速装置，所述TCP加速装置分别与网络数据分析功能NWDAF、用户面功能UPF和外部网关通信连接，所述方法包括：
    向所述NWDAF发送TCP流调整策略订阅指令；
    接收所述NWDAF根据所述TCP流量调整策略订阅指令反馈的TCP流量调整策略，所述TCP流量调整策略由所述NWDAF基于会话管理功能SMF反馈的SMF数据和所述UPF反馈的UPF数据确定；
    根据所述TCP流量调整策略，对各应用程序APP访问网络时产生的TCP流进行调整。
  8. 如权利要求7所述的TCP流的调整方法，其中，所述根据所述TCP流量调整策略，对各应用程序APP访问网络时产生的TCP流进行调整，包括：
    从所述TCP流调整策略中读取需要进行加速的目的IP地址和需要进行限速的目的IP地址；
    根据所述需要进行加速的目的IP地址，对所述需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速；
    根据所述需要进行限速的目的IP地址，对所述需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速。
  9. 如权利要求8所述的TCP流的调整方法，其中，在所述从所述TCP流调整策略中读取需要进行加速的目的IP地址和需要进行限速的目的IP地址之前，所述方法还包括：
    判断所述TCP流调整策略中是否携带了所述需要进行加速的目的IP地址和所述需要进行限速的目的IP地址；
    若携带了所述需要进行加速的目的IP地址和所述需要进行限速的目的IP地址，则执行所述根据所述需要进行加速的目的IP地址，对所述需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速，以及所述根据所述需要进行限速的目的IP地址，对所述需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速的步骤；
    若仅携带了所述需要进行加速的目的IP地址，则仅执行所述根据所述需要进行加速的目的IP地址，对所述需要进行加速的目的IP地址对应的APP访问网络时产生的TCP流进行加速的步骤；
    若仅携带了所述需要进行限速的目的IP地址，则仅执行所述根据所述需要进行限速的目的IP地址，对所述需要进行限速的目的IP地址对应的APP访问网络时产生的TCP流进行限速的步骤。
  10. 一种TCP流的调整装置，包括：
    日志订阅模块，用于分别向会话管理功能SMF和用户面功能UPF发送日志订阅指令；
    数据接收模块，用于接收所述SMF根据所述日志订阅指令反馈的SMF数据，以及所述UPF根据所述日志订阅指令反馈的UDF数据；
    数据分析模块，用于对所述SMF数据和所述UDF数据进行分析，预估TCP流调整策略；
    策略推送模块，用于将所述TCP流调整策略推送至预先接入所述UPF与外部网关之间的TCP加速装置，供所述TCP加速装置根据所述TCP流调整策略对各应用程序APP访问网络时产生的TCP流进行调整。
  11. 一种TCP流的调整装置，包括：
    策略订阅模块，用于向所述NWDAF发送TCP流调整策略订阅指令；
    策略接收模块，用于接收所述NWDAF根据所述TCP流量调整策略订阅指令反馈的TCP流量调整策略，所述TCP流量调整策略由所述NWDAF基于会话管理功能SMF反馈的SMF数据和所述UPF反馈的UPF数据确定；
    TCP流调整模块，用于根据所述TCP流量调整策略，对各应用程序APP访问网络时产生的TCP流进行调整。
  12. 一种网络数据分析功能，包括：
    至少一个处理器；以及，
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行如权利要求1至6中任一项所述的TCP流的调整方法。
  13. 一种TCP加速装置，包括：
    至少一个处理器；以及，
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行如权利要求7至9中任一项所述的TCP流的调整方法。
  14. 一种计算机可读存储介质，存储有计算机程序，所述计算机程序被处理器执行时实现权利要求1至6中任一项所述的TCP流的调整方法，或者权利要求7至9中任一项所述的TCP流的调整方法。
  15. 一种TCP流的调整***，包括：
    如权利要求13所述的TCP加速装置；以及，
    与所述TCP加速装置通信连接的如权利要求12所述的网络数据分析功能NWDAF、用户面功能UPF和外部网关；以及，
    与所述NWDAF通信连接的会话管理功能SMF；其中，
    所述NWDAF还与所述UPF通信连接；
    所述NWDAF用于执行如权利要求1至6任一项所述的TCP流的调整方法；
    所述TCP加速装置用于执行如权利要求7至9任一项所述的TCP流的调整方法。
PCT/CN2021/121789 2020-12-04 2021-09-29 Tcp流的调整方法和*** WO2022116665A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011408143.8A CN114666423A (zh) 2020-12-04 2020-12-04 Tcp流的调整方法和***
CN202011408143.8 2020-12-04

Publications (1)

Publication Number Publication Date
WO2022116665A1 true WO2022116665A1 (zh) 2022-06-09

Family

ID=81853802

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/121789 WO2022116665A1 (zh) 2020-12-04 2021-09-29 Tcp流的调整方法和***

Country Status (2)

Country Link
CN (1) CN114666423A (zh)
WO (1) WO2022116665A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115102884A (zh) * 2022-06-23 2022-09-23 青岛联众芯云科技有限公司 一种用于工控机应用程序的远程数据流量统计方法及装置
CN116095750A (zh) * 2023-01-13 2023-05-09 广州爱浦路网络技术有限公司 数据面转发方法、装置、电子设备及可读存储介质
CN116545942A (zh) * 2023-07-04 2023-08-04 腾讯科技（深圳）有限公司 数据传输方法、装置、电子设备和存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170332282A1 (en) * 2016-05-13 2017-11-16 Huawei Technologies Co., Ltd. Method and system for providing guaranteed quality of service and quality of experience channel
CN110166377A (zh) * 2018-02-14 2019-08-23 华为技术有限公司 一种资源分配方法和装置
CN110300006A (zh) * 2018-03-21 2019-10-01 ***通信有限公司研究院 数据处理方法及装置、功能实体及存储介质
CN110972200A (zh) * 2018-09-30 2020-04-07 华为技术有限公司 通信方法和相关设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170332282A1 (en) * 2016-05-13 2017-11-16 Huawei Technologies Co., Ltd. Method and system for providing guaranteed quality of service and quality of experience channel
CN110166377A (zh) * 2018-02-14 2019-08-23 华为技术有限公司 一种资源分配方法和装置
CN110300006A (zh) * 2018-03-21 2019-10-01 ***通信有限公司研究院 数据处理方法及装置、功能实体及存储介质
CN110972200A (zh) * 2018-09-30 2020-04-07 华为技术有限公司 通信方法和相关设备

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
HUAWEI, HISILICON, INTEL, CONVIDA WIRELESS LLC, TELECOM ITALIA, CHINA MOBILE, CHINA UNICOM, CATR, ALIBABA, NEC, AT&T, ETRI, CA: "Solution for Key Issue 5: NWDAF-Assisted QoS Profile Provisioning", 3GPP DRAFT; DRAFT S2-185706_WAS5177_WAS3635_SOLUTION FOR KEY ISSUE 5 NWDAF-ASSISTED QOS PROFILE PROVISIONING_ERICSSON-HW-LTH-R2-CLEAN, 3RD GENERATION PARTNERSHIP PROJECT (3GPP), MOBILE COMPETENCE CENTRE ; 650, ROUTE DES LUCIOLES ; F-06921 SOPHIA-ANTI, vol. SA WG2, no. Sanya, China; 20180416 - 20180420, 31 May 2018 (2018-05-31), Mobile Competence Centre ; 650, route des Lucioles ; F-06921 Sophia-Antipolis Cedex ; France , XP051459038 *
HUAWEI, HISILICON, INTEL, CONVIDA WIRELESS LLC: "Solution for QoS Provisioning and Adjustment", 3GPP DRAFT; S2-183635_SOLUTION FOR QOS PROVISIONING AND ADJUSTMENT, 3RD GENERATION PARTNERSHIP PROJECT (3GPP), MOBILE COMPETENCE CENTRE ; 650, ROUTE DES LUCIOLES ; F-06921 SOPHIA-ANTIPOLIS CEDEX ; FRANCE, vol. SA WG2, no. Sanya, China; 20180416 - 20180420, 10 April 2018 (2018-04-10), Mobile Competence Centre ; 650, route des Lucioles ; F-06921 Sophia-Antipolis Cedex ; France , XP051437936 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115102884A (zh) * 2022-06-23 2022-09-23 青岛联众芯云科技有限公司 一种用于工控机应用程序的远程数据流量统计方法及装置
CN115102884B (zh) * 2022-06-23 2023-07-21 青岛联众芯云科技有限公司 一种用于工控机应用程序的远程数据流量统计方法及装置
CN116095750A (zh) * 2023-01-13 2023-05-09 广州爱浦路网络技术有限公司 数据面转发方法、装置、电子设备及可读存储介质
CN116095750B (zh) * 2023-01-13 2023-10-31 广州爱浦路网络技术有限公司 数据面转发方法、装置、电子设备及可读存储介质
CN116545942A (zh) * 2023-07-04 2023-08-04 腾讯科技（深圳）有限公司 数据传输方法、装置、电子设备和存储介质
CN116545942B (zh) * 2023-07-04 2023-09-26 腾讯科技（深圳）有限公司 数据传输方法、装置、电子设备和存储介质

Also Published As

Publication number Publication date
CN114666423A (zh) 2022-06-24

Similar Documents

Publication Publication Date Title
WO2022116665A1 (zh) Tcp流的调整方法和***
US11758416B2 (en) System and method of network policy optimization
CA3112926C (en) Slice information processing method and apparatus
US10091644B2 (en) System and method for implementing capability exposure, and capability exposure platform
US20210282053A1 (en) Transfer policy determining method, and apparatus
WO2021155579A1 (zh) 一种数据分析方法、装置及***
CN112840693A (zh) 利用5g移动网络***中的网络分析信息的高效mico模式管理方法
JP2022509188A (ja) ユーザアクセス制御方法、情報送信方法および装置
Moscholios et al. State-dependent bandwidth sharing policies for wireless multirate loss networks
CN115633380B (zh) 一种考虑动态拓扑的多边缘服务缓存调度方法和***
US20230403223A1 (en) Data analysis apparatus management and control method and communication apparatus
US20230037402A1 (en) Communication method, apparatus, and system
WO2022116061A1 (zh) 一种通信方法和装置
WO2022001480A1 (zh) 热门应用识别方法、网络***、网络设备及存储介质
CN116867108A (zh) 一种重构用户面功能网元关键服务能力的方法及通信装置
CN115484621A (zh) 策略优化方法、装置、电子设备及存储介质
WO2024113530A1 (zh) 传输信息的分析方法、装置、设备及存储介质
WO2023185807A1 (zh) 一种通信方法及装置
CN114640981B (zh) 处理5g核心网巨链接的优化方法及装置、设备
WO2024104339A1 (zh) 数据传输方法、装置、相关设备及存储介质
WO2023169101A1 (zh) 一种通信方法和通信装置
US20240129876A1 (en) Systems and methods for analytics and information sharing between a radio access network and a core network
WO2021197463A1 (zh) 节能指示方法、装置及计算机存储介质
WO2023207870A1 (zh) 数据处理方法、终端及网络侧设备
WO2023125211A1 (zh) 一种通信方法及装置

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 03/11/2023)

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

Ref document number: 21899699

Country of ref document: EP

Kind code of ref document: A1