UPF, UPF graceful reboot method, and UPF graceful reboot program

Abstract

To provide an UPF (Unified Protection Framework) for avoiding congestion at the time of a large-scale failure by mainly suppressing communication related to registration.SOLUTION: The UPF210 includes a destination IP address statistical acquisition unit 211 that acquires a set X of destination IP addresses to be communicated at the time of terminal connection, a graceful-restart execution instruction reception unit 212 that receives an instruction to execute a graceful restart, a GTP packet acquisition unit 213 that acquires GTP packets, a cluster division unit 214 that divides a set of SIPM of the GTP packets into N clusters, an on-flag addition unit 215 that adds on-flags to the N clusters at regular time intervals, and a packet transmission control unit 216 that transmits packets whose destination IP addresses belong to a range other than the range X and packets whose destination IP addresses belong to the range X and whose SIPM belongs to a cluster to which an on-flag is added to the Internet side.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a graceful reboot of a User Plane Function (UPF). [Background technology]

[0002] Fifth-generation mobile communication systems (5G) are becoming increasingly popular as a mobile communication technology. However, 5G core network equipment may experience congestion from terminals when restarting, making it unable to restart properly. The communication protocol used in 5G is TCP / IP (Transmission Control Protocol / Internet Protocol). The outline of congestion control by TCP is to reduce the amount of data that a sending terminal sends at one time (this amount is called the "congestion window size") in response to the detection of a congestion state. Details of congestion control by TCP are explained in Non-Patent Document 1. TCP-based congestion avoidance technology repeatedly retries, such as returning to the beginning, unless a certain sequence such as authentication is completed, which can further worsen congestion and exacerbate failures. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Yu Nakayama, "1st What is TCP Congestion Control?" [online], July 1, 2019, [Retrieved July 19, 2024], Internet,<https: / / gihyo.jp / admin / serial / 01 / tcp-cc / 0001> Summary of the Invention [Problem to be solved by the invention]

[0004] 5G core network equipment may not be able to restart properly due to congestion from terminals when restarting.

[0005] The present invention aims to avoid congestion from terminals and enable restart of 5G core network equipment. [Means for solving the problem]

[0006] The inventors discovered that communication related to registration can be suppressed mainly by using the IMSI of the UPF used in line authentication or the TMSI that is the basis of the IMSI, and have completed the present invention.

[0007] (1) The UPF of the present invention includes a terminal-connection destination IP address statistics acquisition unit that acquires statistics on the destination IP addresses of packets communicated when a terminal is connected and designates the set of acquired destination IP addresses as "destination IP address range X," a graceful reboot execution instruction receiving unit that receives an instruction to execute a graceful reboot, a GTP packet acquisition unit that acquires GTP packets from a core network, a cluster division unit that divides SIMs into multiple clusters based on the IMSI or TMSI linked to the GTP packets, an open flag assignment unit that sequentially assigns an open flag to the multiple clusters at regular time intervals, and a packet transmission control unit that sends all packets whose destination IP addresses belong to a range other than range X to the Internet side, and sends only packets of SIMs whose destination IP addresses belong to range X to the Internet side.

[0008] (2) A UPF graceful reboot method according to the present invention is executed by a computer through the steps of: collecting statistics on the destination IP addresses of packets communicated when a terminal is connected; and designating the set of acquired destination IP addresses as "destination IP address range X"; accepting an instruction to execute a graceful reboot; acquiring GTP packets from a core network; dividing SIMs into multiple clusters based on the IMSI or TMSI associated with the GTP packets; sequentially assigning an open flag to the multiple clusters at regular time intervals; and sending all packets whose destination IP addresses belong to a range other than range X to the Internet, and sending only packets of SIMs whose destination IP addresses belong to range X to the Internet.

[0009] (3) The UPF graceful reboot program of the present invention is for causing a computer to function as the UPF. [Effects of the Invention]

[0010] According to the present invention, the cause of congestion can be eliminated by passing communication between each terminal (SIM) and communication destination (destination IP address). [Brief explanation of the drawings]

[0011] [Figure 1] A diagram showing the architecture of a 5G system. [Figure 2] FIG. 2 is a functional block diagram of a UPF according to an embodiment of the present invention. [Figure 3] FIG. 1 is a diagram showing a process flow diagram according to an embodiment of the present invention. [Figure 4] FIG. 10 is a diagram schematically illustrating a state after a set of SIMs has been divided by a cluster dividing unit according to an embodiment of the present invention. [Figure 5] FIG. 10 is a diagram showing the change over time in the number of packets sent from the UPF to the Internet side in the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a diagram illustrating a 5G system architecture 100. UE (User Equipment) 180 is a terminal. (R)AN (Radio Access Network) 190 is a (radio) access network. The UPF (User Plane Function) 210 is a function that transfers user data packets. An International Mobile Subscriber Identity (IMSI) or a Temporary Mobile Subscriber Identity (TMSI) is stored in a packet that arrives at the UPF 210 from the UE 180 side.

[0013] The IMS (IP Multimedia Subsystem) 230 provides services using various media such as video and audio in a communications carrier's network that is entirely IP-based. The VoLTE (Voice over Long Term Evolution) server 240 is a server that provides voice calls as data communications (packet communications). When a terminal connects to a cellular network, it registers with the VoLTE server 240. The SMF (Session Management Function) 150 performs session management.

[0014] FIG. 2 is a functional block diagram of the UPF 210 according to the embodiment of the present invention. The UPF 210 includes a terminal connection destination IP address statistics acquisition unit 211 , a graceful reboot execution instruction reception unit 212 , a GTP packet acquisition unit 213 , a cluster division unit 214 , an open flag assignment unit 215 , and a packet transmission control unit 216 . The terminal connection destination IP address statistics acquisition unit 211 acquires statistics of the destination IP addresses of packets communicated when the terminal is connected, and defines the set of acquired destination IP addresses as "destination IP address range X." The graceful reboot execution instruction receiving unit 212 receives an instruction to execute a graceful reboot from the SMF 150, for example. The GTP packet acquisition unit 213 acquires a GTP (General Packet Radio Service Tunneling Protocol) packet from the core network. The cluster division unit 214 divides the SIM (Subscriber Identity Module) into N clusters based on the IMSI or TMSI associated with the GTP packet. As a method for dividing into clusters, for example, there is a method of dividing according to a remainder or a checksum remainder.

[0015] The open flag assigning unit 215 assigns open flags to N clusters in sequence at regular time intervals. Of packets whose destination IP addresses fall within range X, packets from SIMs belonging to clusters with open flags assigned are sent to the Internet, but packets from SIMs belonging to clusters without open flags assigned are discarded without being sent to the Internet. As time passes, the number of clusters with open flags assigned increases. The packet transmission control unit 216 transmits to the Internet all packets whose destination IP addresses are outside of range X. Then, for packets whose destination IP addresses are within range X, it transmits to the Internet only packets of SIMs that belong to clusters to which an open flag has been assigned.

[0016] FIG. 3 is a process flow diagram of an embodiment of the present invention. In step S302, terminal connection destination IP address statistics acquisition unit 211 takes statistics of destination IP addresses that communicate when the terminal is connected. Then, the destination IP addresses for which statistics have been taken are set as "destination IP address range X." In step S304, for example, the SMF 150 instructs the UPF 210 to execute a graceful reboot. The graceful reboot execution instruction receiving unit 212 receives the instruction to execute a graceful reboot.

[0017] In step S306, the GTP packet acquisition unit 213 acquires the GTP packet from the core network. In step S308, the cluster dividing unit 214 divides the SIM into N clusters based on the IMSI or TMSI associated with the GTP packet.

[0018] FIG. 4 is a diagram showing a schematic diagram of the state after the cluster dividing unit 214 divides the set of SIMs. The SIMs of packets whose destination IP addresses belong to range X are divided into N clusters. The SIMs of packets whose destination IP addresses belong to a range other than range X are also divided into N clusters, but packets whose destination IP addresses belong to a range other than range X are not processed based on clusters in later steps, so they are shown as a single cluster in Figure 4.

[0019] In step S310, the open flag assigning unit 215 assigns open flags to N clusters in sequence at regular time intervals. Therefore, as time passes, the number of clusters to which open flags are assigned increases. In step S312, the packet sending control unit 216 sends to the Internet all packets whose destination IP addresses are outside of range X. For packets whose destination IP addresses are within range X, the packet sending control unit 216 sends only packets of SIMs belonging to clusters to which an open flag has been assigned.

[0020] FIG. 5 is a diagram showing the change over time in the number of packets sent from the UPF 210 to the Internet side. Initially, only packets with destination IP addresses outside range X and packets with destination IP addresses within range X and cluster 0 are sent out. As the number of clusters with open flags increases every fixed time Δt, the number of packets sent out to the Internet increases in a stepwise manner.

[0021] For example, the packets sent in each time interval are as follows: Time interval t0 to t1: Outside range X, and range X and cluster 0 t1~t2: Outside range X, and range X and cluster 0~1 t2~t3: Outside range X, and range X and cluster 0~2 t3~t4: Outside range X, and range X and cluster 0~3 t4~t5: Outside range X, and range X and cluster 0~4 t5~t6: Outside range X, and range X and cluster 0~5

[0022] In step S302, when terminal-connection-time destination IP address statistics acquisition unit 211 takes statistics of destination IP addresses with which the terminal communicates when connected, destination IP address range X includes devices that register mostly when the terminal is connected to a cellular network, such as VoLTE server 240. Outside range X, servers that are not accessed intensively and are heavily affected by retries, such as general servers, are included.

[0023] In this embodiment, immediately after the graceful reboot begins, packet transmission to devices that perform most of the registrations when the terminal is connected to a cellular network is restricted, which reduces the load on the registration process and prevents the terminal from going down due to a large number of registrations being unable to be made, thereby enabling recovery while avoiding congestion. [Quantitative design method 1]

[0024] In practical operation, the values ​​are set as follows: Set the number of terminals in one cluster so that the target server can process transactions of active terminals in one cluster within the expected retry interval. In other words, set the number of terminals belonging to one cluster as [Maximum processing rate of target server] x [Expected retry interval] ÷ [Expected terminal up rate (percentage of active target terminals)]. As an example, if you calculate [maximum processing rate of the target server (10,000 tps)] x [expected retry interval (30 seconds)] ÷ [expected terminal up rate (percentage of target terminals that are running) (0.5)], you get 600,000 units. Note that tps (Transactions per second) is the number of transactions that the target server completes processing in one second.

[0025] Naturally, the number of clusters N can be calculated by dividing the total number of devices by the number of devices belonging to one cluster. For example, if the total number of devices is 30 million, then the number of clusters N is 50.

[0026] Set the time interval for opening a cluster to [Expected retry interval in seconds] + [Number of seconds required for registration]. As an example, if you calculate [expected retry interval in seconds (30 seconds)] + [number of seconds required for registration (10 seconds)], the result is 40 seconds.

[0027] The time required to open all clusters is calculated by multiplying the time interval between cluster openings by the number of clusters, N. In one example, this is calculated as [time interval for opening clusters (40 seconds)] x [number of clusters N (50)], which is approximately 34 minutes. In other words, if there are 30 million devices, recovery would take approximately 34 minutes. In the case of this quantitative design method 1, it is only necessary to ask the target server for the "maximum processing rate" and the "expected retry interval," and it is possible to determine the recovery method in a short time. [Quantitative design method 2]

[0028] In quantitative design method 1, the "expected retry interval" is obtained from the target server. However, the "expected retry interval" can be set from other quantities. All terminals are divided into an appropriate number of clusters, and for one of these clusters, the number of terminals that communicate with each other within a certain period of time is counted.

[0029] The "expected retry interval" can be set by multiplying the number of terminals belonging to one cluster by the given time period and dividing the number of terminals that communicated with the terminal during the given time period. As an example, if we calculate [number of terminals belonging to one cluster (600,000)] x [the above-mentioned fixed time (60 seconds)] ÷ [number of terminals that communicated from the terminal during the above-mentioned fixed time (200,000)], we get 180 seconds. According to this method, it is only necessary to know the "maximum processing rate" from the target server. [Effect 1]

[0030] This embodiment uses the IMSI of the UPF used in line authentication or the TMSI that is the basis for it to suppress mainly registration-related communications, thereby avoiding congestion that is likely to occur when recovering from a large-scale failure. [Effect 2]

[0031] There are two types of network failures: system failures and communication failures. A system failure occurs, for example, when the database logic goes wrong. A communication failure occurs when, for example, the database is operating normally, but the amount of communication is so large that it cannot be handled. In the case of a system failure, it is difficult to operate the recovery function correctly for software that has fallen into an abnormal state under an excessively high load. In this case, in most systems, recovery from the abnormal state is possible by performing a hard reboot, which involves the loss of sessions, but recovery by other methods is often difficult.

[0032] As such, performing a hard reboot is the most effective way to recover from a system failure. However, if a hard reboot is performed, it is certain that the failure will affect all terminals at that point, and in most cases, the number of affected terminals will clearly exceed the threshold. Furthermore, performing a hard reboot normally takes more than an hour to recover, and there is a high possibility of a serious accident. For these reasons, it is necessary to be cautious when performing a hard reboot in the event of a system failure. In a communication system, if a hard reboot is performed, most terminals will try to resume communication after the hard reboot, which will almost certainly result in recovery accompanied by congestion, and there is a risk that recovery will become impossible. Therefore, it is necessary to avoid hard reboots as much as possible. These circumstances are one of the factors that hinder recovery in the event of a serious accident.

[0033] However, if the method of this embodiment is used, it is predicted that sessions with all terminals can be started within one hour even in the event of a hard reboot (according to the quantitative design method 1 described above, recovery is possible in about 34 minutes for 30 million terminals). By using the method of this embodiment, even in situations where the cause is unknown, it is possible to carry out a hard reboot at a relatively early stage after a failure occurs and recover the system while avoiding a serious accident. [Effect 3]

[0034] Another important point is that in this embodiment, packets sent from the terminal to the Internet are discarded, but packets sent from the Internet to the terminal are not discarded. This has the effect of preventing the server from being burdened by starting a TCP session, even in a situation where the terminal immediately retries, putting the terminal into a state of waiting for a response and effectively extending the retry interval.

[0035] This will, for example, make it possible to avoid congestion when restarting core network equipment for the 5th generation mobile communications system, thereby contributing to Goal 9: "Build resilient infrastructure, promote sustainable industrialization and foster innovation."

[0036] Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments. Furthermore, the effects described in the above-described embodiments are merely a list of the most preferable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the embodiments.

[0037] The UPF is implemented by software. When implemented by software, the programs that make up this software are installed on an information processing device (computer). These programs may be recorded on removable media such as CD-ROMs and distributed to users, or may be distributed by being downloaded to users' computers via a network. [Explanation of symbols]

[0038] 100 5G System Architecture 150 SMF 180 UE 190 (R)AN 210 UPF 211 Destination IP address statistics acquisition unit when terminal is connected 212 Graceful reboot execution instruction reception unit 213 GTP packet acquisition unit 214 Cluster division part 215 open flag assignment part 216 Packet sending control unit 220DN 230 IMS 240 VoLTE Server

Claims

1. a terminal-connection-time destination IP address statistics acquisition unit that acquires statistics of destination IP addresses of packets communicated when the terminal is connected, and defines a set of acquired destination IP addresses as a "destination IP address range X"; a graceful reboot execution instruction receiving unit that receives an instruction to execute a graceful reboot; a GTP packet acquisition unit that acquires GTP packets from a core network; a cluster division unit that divides the SIM into a plurality of clusters based on the IMSI or TMSI associated with the GTP packet; an open flag assigning unit that assigns an open flag to a plurality of clusters at regular time intervals; a packet transmission control unit that transmits all packets whose destination IP addresses are outside the range X to the Internet, and transmits only packets of SIMs belonging to a cluster to which an open flag has been added to packets whose destination IP addresses are within the range X to the Internet; UPF equipped with.

2. a step of collecting statistics of destination IP addresses of packets communicated when the terminal is connected, and defining the set of acquired destination IP addresses as "destination IP address range X"; receiving an instruction to perform a graceful reboot; obtaining a GTP packet from a core network; Dividing the SIM into a plurality of clusters based on the IMSI or TMSI associated with the GTP packets; a step of sequentially assigning an open flag to a plurality of clusters at regular time intervals; sending all packets whose destination IP addresses are outside the range X to the Internet, and sending only packets of SIMs belonging to a cluster to which an open flag has been assigned to packets whose destination IP addresses are within the range X to the Internet; The UPF graceful reboot method is performed by the computer.

3. A UPF graceful reboot program for causing a computer to function as the UPF according to claim 1.

Images