Keywords: 5G, edge computing, edge technology session and business continuity
1. The status quo and problems of 4G scenarios
In the 4G scenario, the mobile communication system network architecture is shown in the following figure:
When a user connects to the Internet through a terminal equipment (UE), in the PDU session creation process, a PDN connection is established between the UE and the APN (Internet or other OTT applications). The user plane path of this PDN connection is a path that includes the UE and the eNB End-to-end user data transmission channels including Uu interface (air interface) between (4G base stations), GTP-U tunnel between eNB and SGW, and GTP-U tunnel between SGW and PGW. With the large-scale movement of user equipment (UE), handover between eNB and SGW may occur, but the PGW as the service anchor remains unchanged until the end of the PDU session, as shown in the following figure:
In this case, all users within the coverage of the PGW can access the APN through this PGW. This PGW device can easily become a performance bottleneck and cannot meet the business needs in high-bandwidth scenarios (such as HD video, VR/AR live broadcast, etc.) .
At the same time, PGW equipment is usually deployed in a centralized manner, deployed in a regional center, and user equipment must access the APN through it. In this case, the transmission path from the user equipment at the edge of the network to the bearer network of the APN is extremely long, and the number of device hops passed by the data packet on the transmission path is extremely large, which will inevitably increase the processing delay and cannot meet the requirements of low delay and high reliability Business requirements in scenarios (such as autonomous driving, Internet of Vehicles, high-precision industrial control, etc.).
2. Improvement of 5G scenario and definition of SSC
In response to the above-mentioned problems in the 4G scenario, the 5G network decided to solve this problem from the initial definition of the overall network architecture. The network architecture of the 5G mobile communication system is shown in the following figure:
Compared with 4G network, the biggest improvement of 5G network is to completely separate the user plane and control plane. As the only user plane network element, UPF can be deployed flexibly according to business requirements. As shown in the above figure, UPF PSA1 can be used as a service anchor deployed in a regional center to meet the basic Internet requirements for wide coverage; UPF PSA2 can be used as a local service anchor deployed at the edge to meet local low-latency and high-reliability scenarios Business needs.
In response to the Session and Service Continuity (Session and Service Continuity) issues caused by the distributed sinking deployment of 5G UPF and the localization of service applications, 3GPP has provided three session and service continuity management modes.
- SSC mode one
The PDU session anchor point UPF selected when the PDU session is established will not change due to the movement of the end user, that is, the user's IP address remains unchanged.
- SSC mode two
When the end user adopts the SSC mode 2, the user leaves the original UPF area, the network will trigger the release of the original PDU session and anchor UPF, and instruct the end user to select a new UPF to establish a new PDU session with the same DN.
- SSC mode three
In SSC mode 3, the network allows the user to maintain the PDU session between the user and the original PDU session anchor before the establishment of a new PDU session (the new PDU session anchor point is connected to the same DN). At this time, the user has two The UPF session anchor and PDU session, and finally release the original PDU session.
It can be seen that the SSC mode 1 is similar to the 4G scenario. As the end user moves, the PDU session anchor point remains unchanged. In SSC mode 2, as the end user moves, the original PDU session anchor is first disconnected, and then a connection is established with the new PDU session anchor. In SSC mode three, as the end user moves, the connection with the new PDU session anchor is first established, and then the connection with the original PDU session anchor is disconnected.
3. Flexible use of SSC in 5G scenarios
In the 5G scenario, when users access different types of applications, different types of SSC modes can be flexibly defined according to business requirements. The parameter of SSC mode is carried when the user initiates the PDU session establishment process triggered by the access application.
For ordinary Internet requirements, application servers are usually deployed in regional centers, and SSC mode one can be adopted, that is, as end users move, UPF always remains unchanged. Such applications do not have particularly high requirements for bandwidth and delay, and application server centralization can meet the requirements. The scene diagram is shown in the figure below:
For high-bandwidth business scenarios (such as high-definition video, VR/AR live broadcast, etc.), application servers are usually deployed in edge DCs, and SSC mode two can be used, that is, as the end user moves, the connection with the original UPF is first disconnected. Then establish a connection with the UPF with the same application server exit (N6). Although such applications have higher requirements for bandwidth, the video playback buffering or freezing caused by a short service interruption will not cause an understanding of the overall content. The scene diagram is shown in the figure below:
For low-latency, high-reliability business scenarios (such as autonomous driving, Internet of Vehicles, high-precision industrial control, etc.), application servers are usually deployed in the local DC, and SSC mode 3 is recommended, that is, as end users move, establish and Connect the UPF with the same application server exit (N6), and then disconnect the original UPF. Such applications have extremely high requirements for delay and reliability, and no service interruption is allowed. The scene diagram is shown in the figure below:
For SSC mode 2 and SSC mode 3, as the user equipment moves, it involves the switching of different edge UPF, MEP, and edge applications. In order to maintain user session and business continuity, it is necessary to consider that the user’s proprietary data and edge applications are in different Synchronization between edge nodes.
The user’s proprietary data and edge applications and other related information can be obtained through the path "OSS->source MEPM->source MEP->source edge application", and then through the path "OSS->destination MEPM->destination MEP->destination edge application" "Migrate user-specific data and edge applications and other related information, and synchronize user data and edge applications before switching (SSC mode 3) or after switching (SSC mode 2), so that user sessions and business continuity can be achieved.
Reference materials:
1. 3GPP TS 23.401
2. 3GPP TS 23.501
3. 3GPP TS 23.502