Main and standby switching system and method for high-throughput satellite system core network user plane network elements

Abstract

The invention provides a system and a method for switching main and standby user plane network elements of a core network of a high-throughput satellite system. The system and the method are used for solving the problem of split brain between main and standby equipment and the problem that the context of a user plane and the control plane network elements are inconsistent after the standby equipment is converted into the main equipment. Comprising an MME (Mobility Management Entity), a Redis database, an XGW (Extensible Gateway) and a watchdog thereof, the MME carries main and standby state attributes through a heartbeat message to specify the main and standby states of XGW equipment, and the XGW receives the indication of the MME to set and switch the main and standby states. And the MME stores a user control plane and a user plane context in a redis library, after the XGW standby is switched to the master, the MME performs access information user plane context recovery on the XGW standby, the master XGW reconstructs the user plane context, and the user service is recovered to be normal. When the MME is started for the first time, all the XGWs are in a standby state, multiple threads initiate heartbeat to all the XGWs at the same time, the XGWs which receive heartbeat responses first by the MME conduct standby-to-main conversion, and the other states are standby.

Description

Technical Field

[0001] The invention belongs to the technical field of high-throughput satellite communications. Background Art

[0002] In high-throughput satellite systems, gateway stations are the main control systems for ground satellite communications. Figure 1 As shown in the figure, the gateway station core network-EPC is the access control and data forwarding unit of the entire satellite system. As we all know, the stable operation of the gateway station of the high-throughput satellite communication system is very important. The core network realizes the access control and user data message forwarding functions. The satellite station will be online for a long time after access, so the stable operation of the data plane network element of the core network plays a vital role in the stability of the entire system.

[0003] The traditional user plane equipment master-slave mechanism adopts negotiation between the master and backup devices or third-party sentinel management, which has the following problems

[0004] 1. When the network environment is unstable, a brain split is likely to occur: all devices are in the master or backup state, causing the system to be unable to operate normally.

[0005] 2. The user plane device saves and restores the user plane context by itself. Since the context may be inconsistent with the context corresponding to the control plane network element, the system may run incorrectly. Summary of the invention

[0006] In view of this, the present invention proposes a method and system for solving the switching of the user plane network elements between the active and standby states of the core network of the high-throughput satellite system. In the present invention, the control plane network element leads the active and standby states of the data plane network elements, the active and standby switching process, data recovery, etc., to solve the problem of brain split between the active and standby devices and the problem of inconsistency between the user plane context and the control plane network element after the standby is switched to the active state.

[0007] The core network of the present invention mainly includes two network elements, MME and XGW. Among them, MME is responsible for control plane related functions, including: S1 link management, small station mobility management, small station access control, session management and other functions. XGW is equivalent to the unified deployment of S-GW+P-GW in LTE EPC, and its main functions include packet routing and forwarding of user plane data.

[0008] After the small station is connected to the network, the stability of XGW is crucial to the stability and continuity of user services. When XGW fails, the small station services of the entire system are disconnected and cannot be reconnected, affecting the user experience. Therefore, XGW active and standby are essential for the entire system. When the active XGW network element fails, it is immediately switched to the standby XGW to restore the normal operation of the system.

[0009] The specific technical solutions are as follows:

[0010] A switching system for the user plane network element master and standby of a core network of a high-throughput satellite system includes an MME, a Redis database, an XGW and a watchdog thereof, wherein the MME and the XGW detect whether the XGW is in a normal operating state through a heartbeat mechanism, and the MME determines that the XGW state is abnormal, performs data recovery, performs a master-slave switching process, and quickly performs data service recovery, wherein,

[0011] MME is responsible for control plane related functions, including: S1 link management, small cell mobility management, small cell access control, and session management;

[0012] Redis stores context information. After the small station access process is completed, the MME saves the small station context information, RB bearer information, and GTPU tunnel information in the Redis database;

[0013] XGW is responsible for packet routing and forwarding of user plane data;

[0014] The watchdog is responsible for monitoring the XGW process to prevent XGW from crashing due to software reasons.

[0015] A method for switching the active and standby network elements of a core network user plane of a high-throughput satellite system. The MME specifies the active and standby states of an XGW device by carrying active and standby state attributes in a heartbeat message, and the XGW receives instructions from the MME to set and switch the active and standby states.

[0016] Furthermore, the MME saves the user control plane and user plane context in the redis database. After the backup XGW becomes the active one, the MME restores its access information user plane context, and the active XGW reconstructs the user plane context, and the user service returns to normal.

[0017] Furthermore, when the MME is started for the first time, all XGWs are in standby state, and multiple threads simultaneously initiate heartbeats to all XGWs. The XGW that the MME receives the heartbeat response first switches from standby to active, and the rest are in standby state.

[0018] Furthermore, the MME sets a mutex lock process for the main XGW thread to prevent brain splitting caused by receiving multiple XGW heartbeat messages at the same time.

[0019] Beneficial Effects

[0020] 1. The present invention adopts the control plane network element to dominate the active and standby status of the data plane network element, the switching process, the user plane context recovery, etc. The architecture and message flow are simple, and the existing network element functions are maximized.

[0021] 2. The present invention sets a mutex lock process for the main XGW thread, thereby solving the problem of brain split between the main and standby XGW devices.

[0022] 3. The MME of the present invention stores the user control plane and user plane context in the redis database. After the XGW is switched to the primary state, the MME recovers the access information user plane context, and the primary XGW reconstructs the user plane context, thereby solving the problem of inconsistency between the user plane context and the control plane network element after the backup state is switched to the primary state. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 High-throughput satellite schematic

[0024] Figure 2 Core network hot standby architecture diagram

[0025] Figure 3 When the MME is powered on, the XGW is not active.

[0026] Figure 4 When MME is powered on, XGW is in active use

[0027] Figure 5 The process of the backup switching to the main function due to the failure of the main XGW DETAILED DESCRIPTION

[0028] This embodiment is described by taking two XGWs as an example, but is not limited to two. The present invention designs a set of methods and systems, which mainly include four aspects:

[0029] 1. Core network XGW active / standby architecture diagram

[0030] like Figure 2 ,The core network XGW hot standby mainly includes MME, Redis database, active and standby XGW and its watchdog.

[0031] MME is responsible for control plane interactions such as small station access process and switching, and controls the establishment, recovery, and release of XGW user plane context.

[0032] Redis stores context information. After the small station access process is completed, the MME saves the small station context information, RB bearer information, and GTPU tunnel information in the Redis database.

[0033] XGW is responsible for user plane context construction, opinion data forwarding, etc.

[0034] Watchdog is responsible for watching over the xgw process to prevent xgw from crashing due to software reasons.

[0035] The heartbeat mechanism between MME and XGW detects whether XGW is in normal operation. MME determines that XGW is in abnormal state, performs data recovery, and performs the main-to-backup process to quickly restore data services.

[0036] 2. Small station access context backup

[0037] After the small station access process is completed, the MME saves the S1 link information, small station context information, RB bearer information, and GTPU tunnel information in the Redis database.

[0038] 3. MME cannot obtain the main scenario when powered on

[0039] When the MME is powered on and restarted and cannot obtain the configuration of the active XGW, it sends handshake requests to XGW-A and XGW-B respectively, and the status indication is slave standby, such as Figure 3 A1 and B1 messages.

[0040] (1) XGW-A and XGW-B reply to the MME with a handshake response message, such as Figure 3 , A2, B2.

[0041] (2) MME sets a thread lock to ensure that only one XGW handshake response message can be received and processed at the same time. The XGW that receives the handshake response message first is the primary one. Figure 3 First receive the handshake response from XGW-B, carry the master indication when initiating a handshake to XGW-B, and complete the XGW master selection process.

[0042] Note: In all the above message flows, MME sets thread locks to ensure that only one XGW handshake response message can be received and processed at the same time to prevent the brain split scenario where two active XGWs are indicated at the same time.

[0043] 4. The MME is powered on to obtain the historical active / standby scenarios

[0044] (1) When the MME is powered on and restarted, it obtains the configuration of the active XGW. Figure 4 Get the XGW master configuration as XGW-B.

[0045] (2) MME sends a request to the standby XGW (such as Figure 4 XGW-A) initiates the confirmation process. Figure 4 A1.

[0046] (3) MME sends a request to the primary XGW (such as Figure 4 XGW-B) carries the master indication when initiating handshake, such as Figure 4 B1.

[0047] (4) MME receives the handshake response message from the active XGW and the standby XGW, and completes the XGW active reconfirmation process, such as Figure 4 A2, B2.

[0048] Exception handling: If the handshake confirmation response from the active XGW is not received in step (4), proceed Figure 5 Shown: The failure of the primary XGW causes the backup to become the primary process.

[0049] Note: In all the above message flows, MME sets thread locks to ensure that only one XGW handshake response message can be received and processed at the same time, to prevent the brain split scenario where two active XGWs are indicated at the same time.

[0050] (1) Figure 5 The A1-A4 and B1-B2 processes are the normal active / standby indication heartbeat message interaction processes.

[0051] (2) MME sends a request to the active XGW (such as Figure 5 XGW-B) initiated three master indication messages B3, B4, and B5 in succession, but did not receive a handshake response from the master XGW.

[0052] (3) MME sends a request to the primary XGW (such as Figure 5 XGW-B) initiates a standby indication handshake message to switch from primary to standby, such as Figure 5 B6.

[0053] (4) MME sends a request to the standby XGW (such as Figure 5 XGW-A) initiates a master indication handshake message to switch from standby to master. Figure 5 A5.

[0054] (5) MME receives the active XGW (such as Figure 5 XGW-A) handshake response message such as Figure 5 A6, complete the XGW standby to master confirmation process.

[0055] (6) MME and primary XGW (such as Figure 5 XGW-A) performs the user plane up and down recovery process, the active XGW reconstructs the user plane context, and the user service returns to normal.

[0056] Note: In all the above message flows, MME sets thread locks to ensure that only one XGW handshake response message can be received and processed at the same time to prevent the brain split scenario where two active XGWs are indicated at the same time.

Claims

1. A switching system for active / standby user plane network elements in a core network of a high-throughput satellite system, characterized in that: It includes MME, Redis database, XGW and its watchdog. The heartbeat mechanism between MME and XGW detects whether XGW is in normal operation. MME determines that XGW is abnormal, performs data recovery, performs the main-to-backup process, and quickly recovers data services. MME is responsible for control plane related functions, including: S1 link management, small cell mobility management, small cell access control, and session management; Redis stores context information. After the small station access process is completed, the MME saves the small station context information, RB bearer information, and GTPU tunnel information in the Redis database; XGW is responsible for packet routing and forwarding of user plane data; The watchdog is responsible for monitoring the XGW process to prevent XGW from crashing due to software reasons.

2. A method for switching between active and standby network elements of a user plane in a core network of a high-throughput satellite system, characterized in that: The MME specifies the active and standby status of the XGW device by carrying the active and standby status attributes in the heartbeat message. The XGW accepts the MME's instructions to set and switch the active and standby status.

3. A method for switching between active and standby user plane network elements in a core network of a high-throughput satellite system according to claim 2, characterized in that: MME saves the user control plane and user plane context in the redis database. After the standby XGW becomes active, MME restores its access information user plane context, and the active XGW reconstructs the user plane context, and the user service returns to normal.

4. A method for switching between active and standby user plane network elements in a core network of a high-throughput satellite system according to claim 2, characterized in that: When MME is started for the first time, all XGWs are in standby state. Multithreading initiates heartbeats to all XGWs simultaneously. The XGW that MME receives the heartbeat response first switches from standby to master, and the rest are in standby state.

5. The method for switching between active and standby user plane network elements of a core network of a high-throughput satellite system according to any one of claims 2 to 4, characterized in that: MME sets the mutex lock processing for the main XGW thread to prevent brain split caused by receiving multiple XGW heartbeat messages at the same time.

Images