Video conference communication control method and device

Through distributed clustering and ZooKeeper architecture in the video conferencing system, role allocation and load balancing strategies are used to solve the uniqueness of signaling servers caused by multi-terminal concurrency and server failure, and the stability and load balancing of the video conferencing system are achieved.

CN118921437APending Publication Date: 2024-11-08BEIJING HONGYUN RONGTONG TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411153556.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-21
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

In the case of concurrent requests by multiple terminals and server failures, existing video conferencing systems may cause multiple working signaling servers to be allocated for the same conference, resulting in inconsistent status and abnormalities, and it is difficult to ensure uniqueness and load balancing.

Method used

Adopting distributed clustering and ZooKeeper architecture, through role allocation and load balancing policies, ensures that a unique work signaling server is assigned to each meeting, and automatically identify and allocate new servers in the event of server failure, ensuring consistency and load balancing.

Benefits of technology

In various scenarios (such as subscriber failure, allocator failure, terminal request concurrency, etc.), it is precisely ensure that the same meeting in the video conference is allocated to the unique work signaling server, ensuring final consistency and load balancing, and avoiding abnormal situations with inconsistent status.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118921437A_ABST
    Figure CN118921437A_ABST
Patent Text Reader

Abstract

The invention discloses a video conference communication control method and device, the method is applied to any server in a distributed cluster, the server has the following role types: a publisher, a distributor and a subscriber, the distributed cluster has a distributed coordination service, and the method comprises the following steps: when the server is started, creating a node directory and competing for the distributor; when the distributor is successfully competed, the distributor monitors the change of a conference directory to determine whether a to-be-distributed conference identifier appears or not, and distributes subscribers to the to-be-distributed conference identifier by using the load information of all servers in the cluster, and the subscribers: obtaining the to-be-distributed conference identifier by monitoring the change of the node directory; and creating a result file into the conference directory according to the to-be-distributed conference identifier so as to update the conference directory, and when the server becomes any role, starting a monitoring function of the role on a specified directory. The invention can accurately ensure that a unique signaling server is allocated to the same conference in the video conference.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of computer communication and video conferencing, and in particular to a video conferencing communication control method and device. Background Art

[0002] This section is intended to provide a background or context to the embodiments of the invention recited in the claims. No description herein is admitted to be prior art by inclusion in this section.

[0003] Video conferencing is a communication technology that allows real-time audio and / or video communication via the Internet or other network connections. The server responsible for managing conference information and synchronizing with terminal status is called a signaling server. Since there are competitive resources in a conference, such as the current host information of the conference, the number of speakers in the conference, etc., it is necessary to assign a unique signaling server to a conference as a decision-making and arbitration center. This unique signaling server can be called a working signaling server. In other words, in any case, even if there is poor communication between the server and the terminal, and / or multiple terminals make concurrent requests, strictly assigning a unique working signaling server to each conference is the key to ensuring conference communication control.

[0004] The existing conference system needs to avoid the situation where multiple signaling routing servers make decisions at the same time in each link as much as possible. The function of the signaling routing server is to ensure that the signaling messages in the conference can be accurately transmitted from the source to the destination. Figure 1 For the processing diagram of the existing video conferencing terminal request, refer to Figure 1Before the terminal accesses the signaling routing server, it first uses the conference number to take the remainder of the number of signaling routing servers. In this way, when the order of the signaling routing server list is known and fixed, the signaling routing server addresses that multiple different terminals of the same conference prefer to access are the same. This approach can ensure that in most cases, only one signaling routing server will allocate its corresponding working signaling server for the same conference. However, for disaster recovery considerations, the terminal must add processing logic for signaling server exceptions. Under the premise that the preferred signaling routing server responds timeout, it will communicate with the alternative signaling routing server. The introduction of this logic makes it possible for multiple signaling routing servers to receive the task of allocating signaling servers for the same conference at the same time when multiple terminals concurrently request the signaling server address. When they query the database for the signaling server information of the conference at the same time, they cannot find it, so they make their own decisions, allocate different working signaling servers, and synchronize the results to the corresponding terminals, which results in a chaotic situation where more than one working signaling server is working for the same conference. In this scenario where multiple terminals in the same conference are split across two or more working signaling servers, they can only be converged to the same signaling server by adding a discovery mechanism later. Although eventual consistency can be achieved, both discovery and convergence take time. During this period, the terminals split across two or more signaling services may not be visible to each other or may be in inconsistent states. Summary of the invention

[0005] The embodiment of the present invention provides a video conference communication control method, which is used to accurately ensure that the same conference in the video conference is assigned a unique working signaling server. The method is applied to any server in a distributed cluster. The server has the following role types: publisher, distributor, subscriber; the distributed cluster has a distributed coordination service, and the distributed coordination service includes multiple components that provide consistency, synchronization and coordination functions for the distributed cluster. The method includes:

[0006] When the server starts, a node directory is created; the node directory is used to update the storage server load information and temporarily store the conference ID that the server is responsible for;

[0007] Competing allocators; creating an allocator directory during the competing allocator process, storing allocator information in the allocator directory; the allocator directory is automatically deleted when the allocator fails;

[0008] When a competing allocator succeeds:

[0009] Monitor whether the conference directory appears with the conference ID to be allocated; the conference directory stores the conference ID and the result file corresponding to the conference ID. The conference ID to be allocated is the conference ID that does not have a corresponding result file in the conference directory. The conference directory is created by the publisher, which is the first server accessed by the terminal. The terminal accesses the conference with the conference ID; the publisher is used to: create the conference directory when the terminal accesses for the first time; receive terminal access, and provide the terminal with the subscriber information corresponding to the conference ID by monitoring the changes in the conference directory;

[0010] When a conference identifier to be allocated appears, the load information of all servers in the cluster is obtained, and the subscriber corresponding to the conference identifier to be allocated is determined by using the load information, and the conference identifier to be allocated is updated to the node directory, so that the subscriber corresponding to the conference identifier to be allocated: the conference identifier to be allocated is obtained by monitoring the changes in the node directory, and a result file is created according to the conference identifier to be allocated to the conference directory, and the conference identifier to be allocated in the node directory is deleted at the same time; wherein the result file carries the conference identifier;

[0011] Among them, when the server becomes any role, the role's monitoring function for the designated directory is started; the designated directory includes the node directory, the conference directory, and the distributor directory; different role servers monitor different designated directories; when the publisher and subscriber monitor and receive the message that the distributor directory is deleted, they start to compete for the distributor; when the subscriber fails, the result file of the failure in the conference directory is automatically deleted.

[0012] The embodiment of the present invention also provides a video conference communication control device, which is used to accurately ensure that the same conference in the video conference is assigned a unique working signaling server. The device is applied to any server in a distributed cluster. The server has the following role types: publisher, distributor, subscriber; the distributed cluster has a distributed coordination service, and the distributed coordination service includes multiple components that provide consistency, synchronization and coordination functions for the distributed cluster. The device includes:

[0013] The startup processing module is used to create a node directory when the server starts; the node directory is used to update the storage server load information and temporarily store the conference identifier that the server is responsible for;

[0014] The distributor competition module is used to compete for distributors; during the distributor competition process, a distributor directory is created, and distributor information is stored in the distributor directory; the distributor directory is automatically deleted when a distributor fails;

[0015] The distributor processing module is used to monitor whether the conference identifier to be allocated appears in the conference directory when the competing distributor succeeds; the conference directory stores the conference identifier and the result file corresponding to the conference identifier. The conference identifier to be allocated is the conference identifier that has no corresponding result file in the conference directory. The conference directory is created by the publisher, and the publisher is the first server accessed by the terminal. The terminal access carries the conference identifier; the publisher is used to: create the conference directory when the terminal accesses for the first time; receive terminal access, and provide the terminal with the subscriber information corresponding to the conference identifier by monitoring the changes in the conference directory; when the conference identifier to be allocated appears, the distributor obtains the load information of all servers in the cluster, and uses the load information to determine the subscriber information corresponding to the conference identifier to be allocated. Reader, update the conference identifier to be allocated to the node directory, so that the subscriber corresponding to the conference identifier to be allocated: obtain the conference identifier to be allocated by monitoring the changes in the node directory, create a result file according to the conference identifier to be allocated to the conference directory, and delete the conference identifier to be allocated in the node directory; wherein the result file carries the conference identifier; wherein, when the server becomes any role, the role's monitoring function for the specified directory is started; the specified directory includes the node directory, the conference directory, and the distributor directory; different role servers monitor different specified directories; when the publisher and subscriber monitor and obtain the message that the distributor directory is deleted, they start to compete for the distributor; when the subscriber fails, the failed result file in the conference directory is automatically deleted.

[0016] An embodiment of the present invention further provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned video conference communication control method when executing the computer program.

[0017] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned video conference communication control method is implemented.

[0018] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the above-mentioned video conference communication control method is implemented.

[0019] In an embodiment of the present invention, for any server in a distributed cluster, when the server is started, a node directory is created; the node directory is used to: update the storage server load information, temporarily store the conference identifier that the server is responsible for; compete for the allocator; create an allocator directory in the process of competing for the allocator, and store the allocator information in the allocator directory; the allocator directory is automatically deleted when the allocator fails; when the competing allocator succeeds: monitor whether the conference identifier to be allocated appears in the conference directory; the conference directory stores the conference identifier and the result file corresponding to the conference identifier, the conference identifier to be allocated is the conference identifier that does not have a corresponding result file in the conference directory, the conference directory is created by the publisher, the publisher is the first server accessed by the terminal, and the terminal accesses the conference identifier; the publisher is used to: create the conference directory when the terminal accesses for the first time; receive the terminal access, and provide the terminal with the corresponding information of the conference identifier by monitoring the changes in the conference directory Subscriber information; when a conference identifier to be allocated appears, the distributor obtains the load information of all servers in the cluster, and uses the load information to determine the subscriber corresponding to the conference identifier to be allocated, and updates the conference identifier to be allocated to the node directory, so that the subscriber corresponding to the conference identifier to be allocated: obtains the conference identifier to be allocated by monitoring the changes in the node directory, creates a result file in the conference directory according to the conference identifier to be allocated, and deletes the conference identifier to be allocated in the node directory; wherein the result file carries the conference identifier; wherein, when the server becomes any role, the role's monitoring function for the specified directory is started; the specified directory includes the node directory, the conference directory, and the distributor directory; servers with different roles monitor different specified directories; when the publisher and subscriber monitor and receive the message that the distributor directory is deleted, they start to compete for the distributor; when the subscriber fails, the failed result file in the conference directory is automatically deleted. By using the video conference communication control method in the embodiment of the present invention to allocate working signaling servers, in various scenarios where subscriber failures, distributor failures, publisher failures, terminal request concurrency, and even poor network communication occur, the uniqueness of the working signaling server allocated to the same conference in the video conference can still be accurately guaranteed, and the routing allocation between different conferences can achieve load balancing as much as possible. When the working signaling server fails, it can automatically identify the anomaly and promptly allocate a suitable and unique new server to the conference undertaken by the failed server. The video conference communication control method is not only applicable to existing conferences in the conference system, but also to newly created conferences through various business means. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0021] Figure 1 A schematic diagram of processing requests for existing video conferencing terminals;

[0022] Figure 2 Schematic diagram of the flow of a video conference communication control method in an embodiment of the present invention;

[0023] Figure 3 A diagram showing a specific example of a video conference communication control method according to an embodiment of the present invention;

[0024] Figure 4 FIG. 4 is another specific example diagram of the video conference communication control method in an embodiment of the present invention;

[0025] Figure 5 FIG. 4 is another specific example diagram of the video conference communication control method in an embodiment of the present invention;

[0026] Figure 6 FIG. 4 is another specific example diagram of the video conference communication control method in an embodiment of the present invention;

[0027] Figure 7 Schematic diagram of a video conference communication control device in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0029] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0030] First, the technical terms involved in the embodiments of the present invention are explained:

[0031] Zookeeper architecture: The open source network architecture Zookeeper manages distributed server nodes through an internal data synchronization mechanism. These nodes are organized in a tree structure. In this tree structure, each node is called a znode. These znodes can contain data and have child nodes, forming a multi-level hierarchy. Zookeeper allows users to store and retrieve data in such a tree directory. Each znode can store not only a small amount of data (the default upper limit is 1MB). Zookeeper is responsible for synchronously copying this tree structure and its data to each node.

[0032] There are two types of znodes: persistent and ephemeral.

[0033] -Persistent nodes: Once created, they will exist unless explicitly deleted. However, there is a type of persistent node with an expiration date, which starts from the time of creation and will be automatically deleted after the expiration date.

[0034] -Ephemeral nodes: Their lifecycle is bound to the session that created them. When the session ends, the nodes will be automatically deleted, and ephemeral nodes cannot have child nodes.

[0035] The design principle of Zookeeper is to build a highly available, strongly consistent distributed coordination service. It achieves this by deploying multiple Zookeeper nodes on different servers (it is usually recommended to deploy an odd number of nodes, such as 3, 5, etc.).

[0036] These Zookeeper nodes together form a cluster, which determines a leader and multiple followers through an election mechanism, and may also have observers in some configurations. The cluster as a whole maintains and manages the logical tree structure (znode hierarchy) to provide services to clients.

[0037] In a Zookeeper cluster, each node (whether Leader, Follower or Observer) runs a Zookeeper service replica, that is, a complete Zookeeper service program. These service replicas work together to maintain the distributed coordination services of the entire cluster, including key functions such as data consistency, fault tolerance, and leader node election. Each node has the ability to process client requests.

[0038] Ideally, the information in the service program replica on each node in the Zookeeper cluster should be completely consistent. This is achieved through the atomic broadcast protocol within Zookeeper, which ensures that all write operations are replicated to all nodes. This process ensures data consistency.

[0039] When a client writes to Zookeeper, the request is sent to the current Leader node. The Leader processes the request and generates a transaction log, and then broadcasts the operation to all Follower nodes. After receiving the transaction, the Follower node applies it and feeds it back to the Leader. Once the Leader receives confirmation from more than half of the nodes (including the Leader itself), it commits the transaction and then notifies all Followers to commit. In this way, even if a Follower node is temporarily unable to receive or apply an update due to network problems or other reasons, it will eventually keep data consistent with other nodes through the synchronization mechanism.

[0040] In the prior art, it is easy for multiple terminals of the same conference to be split to two or more signaling servers, which will lead to an abnormal state where the terminals are invisible to each other or the states are inconsistent. Therefore, an embodiment of the present invention provides a video conference communication control method. Figure 2 FIG. 1 is a flow chart of a video conference communication control method according to an embodiment of the present invention. Figure 2 As shown, the method is applied to any server in a distributed cluster. The server has the following role types: publisher, distributor, subscriber. Different roles perform different functions. In the embodiment of the present invention, the same server can have multiple roles at the same time. The distributed cluster has a distributed coordination service. The distributed coordination service includes multiple components that provide consistency, synchronization and coordination functions for the distributed cluster. Through the distributed coordination service, any role can operate the specified data in the cluster (for example, multiple directories in the embodiment of the present invention) through the server (which can be understood as a client). The method includes:

[0041] Step 201: When the server is started, a node directory is created; the node directory is used to update and store server load information and temporarily store conference identifiers that the server is responsible for;

[0042] Step 202, competing for allocators; creating an allocator directory in the allocator competition process, storing allocator information in the allocator directory; the allocator directory is automatically deleted when the allocator fails;

[0043] Step 203, when the competing allocator succeeds: monitor whether the conference identifier to be allocated appears in the conference directory; the conference directory stores the conference identifier and the result file corresponding to the conference identifier, and the conference directory is created by the publisher, and the publisher is the first server accessed by the terminal; the publisher is used to: receive terminal access, and provide the terminal with subscriber information corresponding to the conference identifier by monitoring changes in the conference directory; when the conference identifier to be allocated appears, the allocator obtains the load information of all servers in the cluster, and uses the load information to determine the subscriber corresponding to the conference identifier to be allocated, and updates the conference identifier to be allocated to the node directory, so that the subscriber corresponding to the conference identifier to be allocated: obtains the conference identifier to be allocated by monitoring changes in the node directory, creates a result file in the conference directory according to the conference identifier to be allocated, and deletes the conference identifier to be allocated in the node directory; wherein the result file carries the conference identifier; wherein, when the server becomes any role, the role's monitoring function for the specified directory is started; the specified directory includes the node directory, the conference directory, and the allocator directory; servers with different roles monitor different specified directories; when the publisher and the subscriber monitor and obtain the message that the allocator directory is deleted, they start competing for the allocator; when the subscriber fails, the failed result file in the conference directory is automatically deleted.

[0044] from Figure 2 It can be seen from the shown process that by allocating working signaling servers through the video conference communication control method in the embodiment of the present invention, in various scenarios where subscriber failure, distributor failure, publisher failure, terminal request concurrency and even poor network communication occur, the uniqueness of the working signaling server allocated to the same conference in the video conference can still be accurately guaranteed, and the routing allocation between different conferences can achieve load balancing as much as possible. In addition, when the working signaling server fails, it can automatically identify the anomaly and promptly allocate a suitable and unique new server for the conference undertaken by the failed server.

[0045] The embodiments of the present invention propose solutions for three practical needs in video conferencing: 1. Whether it is an existing conference number or a newly created conference number, a unique working signaling server corresponding to the conference number is allocated; 2. The allocation process refers to the server load information, and different conference numbers are load-balanced as much as possible; 3. After a server failure, a new round of allocation logic is automatically triggered, and the same conference number can still be accurately allocated to a unique working signaling server, and different conference numbers are load-balanced as much as possible. The method adopted by the embodiment of the present invention is to first decide on a unique allocator based on a distributed cluster and distributed coordination service, and then the allocator allocates the corresponding working signaling server to the conference number, schedules the nodes according to the load balancing strategy, and uses the distributed coordination service again during the allocation process to synchronize information between nodes. Among them, the distributed coordination service includes an atomic broadcast coordination function.

[0046] Distributed coordination services are key components that provide consistency, synchronization, and coordination mechanisms in distributed systems. In a distributed system consisting of multiple nodes, these services ensure that all nodes can work together to maintain system stability and consistency even in the face of network partitions, node failures, or other challenges unique to distributed systems. Common distributed coordination services include ZooKeeper architecture, etcd distributed key-value storage, Consul distributed architecture, etc. In the embodiments of the present invention, the distributed coordination service is illustrated by using the ZooKeeper architecture.

[0047] The conventional approach to using Zookeeper to solve consistency problems in business is to agree in advance on one or several fixed business topics (the conference identifier in the corresponding video conference includes the conference number in the embodiment of the present invention), and multiple business server nodes compete to create the same znode as Zookeeper clients. However, although the business server selected by directly using Zookeeper for decision-making can ensure the uniqueness of the conference, the allocation result cannot refer to the load information of the business server. In addition, this method also requires that the business topic be prepared in advance. For the typical scenario in which a terminal temporarily creates and participates in a video conference, the conference number is dynamically generated and cannot be agreed or known in advance by each business server.

[0048] The specific approach is to divide the servers into three types according to the roles they play in the cluster:

[0049] (1) Publisher

[0050] The publisher is the first server that the terminal directly accesses to obtain the working signaling server information corresponding to the conference.

[0051] (2) Allocator

[0052] Servers compete to become distributors through Zookeeper. The distributor is unique in a server cluster and is responsible for allocating work signaling servers for all conferences.

[0053] (3) Subscribers

[0054] A subscriber is a working signaling server for a conference ID.

[0055] It is also necessary to introduce the directory settings in the embodiment of the present invention. Zookeeper is used in the embodiment of the present invention to create nodes, update node data, and monitor node or sub-node changes. The node is similar to a directory or file in a file system and is stored on a Zookeeper server. To avoid ambiguity, it is distinguished from the signaling server node mentioned above, and for the convenience of description and understanding, all references to the concept of nodes in Zookeeper are replaced by directories or files. There are three types of Zookeeper directories: temporary directories, persistent directories, and persistent directories with validity periods. The biggest difference between them is the timing of deletion: temporary directories will be automatically deleted by Zookeeper at the end of the creator session; persistent directories are not related to the life cycle of the creator session and can only be deleted explicitly; persistent directories with validity periods will be automatically deleted by Zookeeper when the set validity period expires.

[0056] The method involved in the embodiment of the present invention uses a temporary directory and a persistent directory with a validity period. Specifically, it includes:

[0057] Category 1: Distributor Directory (created by the distributor)

[0058] (1) / group / router /

[0059] The type is a temporary directory, which contains a temporary file master, which is used to compete for the distributor and store the communication and identity information of the distributor server. group is the name of the group that all servers in the conference system join.

[0060] For example, the server with node ID 1 becomes the group distributor of the group name stp. The following file will be generated on the Zookeeper server: / stp / router / master, which is created by the server with node ID 1 and stores the distributor information, for example: 1#10.130.93.107:8000.

[0061] Category 2: Node Directory (Created by Subscribers)

[0062] (2) / group / subscribers /

[0063] Assume that the group name is stp, and take the server with node ID 1 as an example. After it joins the group, two files will actually be generated on the Zookeeper server:

[0064] The first temporary file: " / stp / subscribers / 1", stores the basic information of the node server and temporarily stores the responsible conference identification information, such as the conference number list.

[0065] The second temporary file: " / stp / subscribers / 1_info" stores node load information, mainly indicators such as the node's CPU / memory / number of connections. The allocator will adopt a load balancing scheduling strategy when allocating signaling servers for specific conferences.

[0066] Both temporary files are in the node directory.

[0067] cpu and memory are the percentage of cpu and memory currently used by the node respectively. The number of connections refers to the number of online terminals, which can be the total number of online users in different conferences.

[0068] Server basic information refers to key details about the server configuration, status, and functions, such as server hardware information, operating system information, software environment information, network configuration information, status report information, system overview information, etc.

[0069] The first temporary file updates the basic information of the storage server in real time and temporarily stores the conference identifier responsible for processing the meeting. For example, when the assignor determines to assign the meeting number 12345 to the server with ID 119, the assignor updates the meeting number 12345 and writes it to the first temporary file so that the server 119 reads the meeting number 12345 and creates the corresponding result file. The second temporary file updates the load information of the storage server in real time. For example, the second temporary file updates the load information of the storage server 119 in real time.

[0070] The first temporary file and the second temporary file are both created by the server corresponding to the node ID, that is, the subscriber, when starting up, and are of temporary file type.

[0071] There are pairs of temporary files created by multiple servers in the node directory.

[0072] When a subscriber fails, the first temporary file and the second temporary file, as well as the result file, created by it are automatically deleted by the Zookeeper service. The distributor can obtain the failed subscriber information by monitoring the changes in the node directory.

[0073] By reading the information in the node directory / group / subscribers, you can get the node information of all servers in the entire group, including communication and load information.

[0074] That is, the basic information of the server and the server load information are permanent information, which will always exist as long as the node does not fail. The conference identifier responsible for processing the conference is temporary information. When the result file corresponding to the conference identifier is created, the conference identifier responsible for processing the conference is automatically deleted.

[0075] Category 3: Conference Directory (created by publisher)

[0076] (3) / group / subjects /

[0077] This directory maintains all conference numbers and allocation results of all signaling servers that need to be allocated in the video conference.

[0078] Assuming that the group name is stp, taking the conference number 123456 as an example, when the first terminal in the conference accesses the server, that is, the publisher, creates a conference directory and monitors the allocation result of the conference ID to be allocated.

[0079] In one embodiment, the conference identifier stored in the conference directory exists in the form of a file named with the conference identifier, and the result file stored in the conference directory exists in the form of a file named with the conference identifier and the first character identifier; the first character is arbitrarily specified;

[0080] The allocator identifies the conference identifier to be allocated by monitoring the conference directory. Specifically, monitoring whether the conference identifier to be allocated appears in the conference directory may include:

[0081] By comparing the file names, the conference ID without result files is determined, and the conference ID without result files is used as the conference ID to be allocated.

[0082] For example, it can be implemented as follows: When the publisher is ready to obtain the address of the working signaling server, it will create the " / stp / subjects / 123456" file on the Zookeeper server and monitor the " / stp / subjects / 123456_Owner" file (i.e., the result file) created by the subscriber. The allocator can obtain the information of all files under the directory " / stp / subjects / " and find that the " / stp / subjects / 123456" file exists, but the " / stp / subjects / 123456_Owner" file does not exist, that is, by comparing the file names, the conference number to be allocated, 123456, is obtained and the allocation logic is started.

[0083] The above second-class node directory and third-class conference directory are used to maintain all node information in the server group, as well as all conference numbers and their corresponding subscriber node information appearing in the conference system (which can be understood as a system with terminals, clusters, distributed coordination services, and multiple program combinations, which can implement the video conference communication control method). Among them, the data of the second-class node directory is the communication and load information of the server node itself, and the type is a temporary directory, so when the server fails, the / group / subscribers / nodeId and / group / subscribers / nodeId_info created by it will be automatically deleted by the Zookeeper service. The / group / subjects / conference number file in the third-class conference directory is created by the server that the terminal directly accesses for the first time, that is, the conference publisher, and the type is a persistent directory with an expiration date, so it will not be deleted when the session between the publisher and the Zookeeper service ends; the / group / subjects / conference number_Owner file in the third-class conference directory is created by the work signaling server corresponding to the allocated conference, that is, the conference subscriber. When the subscriber signaling server fails, the result file will be automatically deleted by the Zookeeper service.

[0084] In the embodiment of the present invention, all created directories support the simultaneous transfer of data during creation.

[0085] When the embodiment of the present invention is implemented, any server joins a group when it is just started. The server may become any one of the publisher, distributor, and subscriber, or any combination of the three roles in one server.

[0086] For any server, when the server starts, it starts to create node directories and compete for distributors. When it is the first server directly accessed by the terminal, it can also become a publisher at the same time.

[0087] When the server becomes any role, the monitoring function of the role on the designated directory is started; the designated directory includes the node directory, the conference directory, and the distributor directory; different role servers monitor different designated directories.

[0088] Specifically, when acting as a distributor, the directories that need to be monitored include but are not limited to the node directory and the conference directory. When acting as a subscriber, the directories that need to be monitored include but are not limited to the node directory and the distributor directory. When acting as a publisher, the directories that need to be monitored include but are not limited to the conference directory and the distributor directory.

[0089] Multiple roles do not conflict with the server. The same server can assume the roles of publisher, distributor, subscriber, or any combination thereof.

[0090] When the competing allocator succeeds, it can also become a publisher and / or a subscriber and perform role functions.

[0091] When the competing allocator fails, it becomes the publisher and / or the subscriber, performing the role functions.

[0092] In one embodiment, competing allocators may include:

[0093] Based on the atomic broadcast coordination function, try to create the distributor directory; the atomic broadcast coordination function is used to ensure that there is only one server in the cluster that can successfully create the distributor directory;

[0094] When the distributor directory is created successfully, you become the distributor.

[0095] For example, multiple servers are started at the same time, trying to create the distributor directory and the master file " / stp / router / master" respectively. The atomic broadcast coordination function ensures that only one server can be created successfully. The other servers will fail to create and get the error code "directory already exists". The distributor directory creation interface supports the simultaneous input of data. The distributor directory data content is its own node information. The distributor directory is created successfully, or the directory already exists but the read directory data is the same as itself, which is considered successful.

[0096] When the competing allocator succeeds, that is, it has successfully created " / stp / router / master" and stored its own allocator node information. After becoming an allocator, it is necessary to always monitor whether there is a conference identifier to be allocated in the conference directory. When it is determined that a conference identifier to be allocated appears, the load information of all servers in the cluster is obtained, and the load information is used to determine the subscriber corresponding to the conference identifier to be allocated. The subscriber is responsible for the relevant processing of the conference number.

[0097] Once the distributor fails, the distributor directory and distributor information master file it created will be automatically deleted. Other servers in the group (including publishers and subscribers) will promptly discover this situation and start competing for new distributors because they monitor the changes in the temporary distributor directory / group / router / master created by the distributor. The newly generated distributor will identify whether there is a new conference ID to be allocated through file name comparison, and if so, it will start executing the allocation logic.

[0098] The publisher is the first server accessed by the terminal. When the publisher fails, the directory created by the publisher is not a temporary directory and will not change. Therefore, a server failure that is neither the distributor nor the subscriber will not have any impact on the conference system.

[0099] When a signaling server that plays the role of a subscriber fails, the relevant conference number result file created by the signaling server in the conference directory is automatically deleted by Zookeeper. The allocator will be notified by monitoring changes in the conference directory and / or monitoring changes in the node directory, and will identify new conference identifiers to be allocated through file name comparison, such as a new list of conference numbers to be allocated, and execute the allocation logic one by one to allocate new subscribers for each conference number. The conference number is updated to the ID file corresponding to the new subscriber, that is, the first temporary file. The new subscriber reads the list of conference numbers for which it is responsible, and then creates a new conference number result file.

[0100] When a signaling server that is both an allocator and a subscriber fails, other signaling servers in the group will first compete for a new allocator, which will then identify a new list of conference numbers to be allocated through file comparison and execute the allocation logic one by one.

[0101] When the three roles are concentrated on one server and the server fails, the other signaling servers in the group will first compete for a new distributor, and then the distributor will identify the new list of conference numbers to be allocated through file comparison and execute the allocation logic one by one. At the same time, the first server to receive the access from the conference terminal is determined as the new publisher.

[0102] In one embodiment, the load information includes CPU usage, memory usage, and number of connections;

[0103] Using the load information to determine the subscriber corresponding to the conference identifier to be allocated may include:

[0104] The CPU usage, memory usage, and number of connections of each server in the cluster are used to determine the subscribers corresponding to the conference ID to be allocated.

[0105] For example, different scores are set for data values ​​of different segments of CPU usage, memory usage, and number of connections, and corresponding weight coefficients are set to obtain the load score value of each server. Based on the load score values ​​of all servers, the optimal server is determined as the subscriber corresponding to the conference identifier to be allocated.

[0106] In the embodiment of the present invention:

[0107] 1. The distributor server must be determined first. The distributor is globally unique. As long as the distributor server is not faulty, no matter which server the terminal looks for to obtain the signaling server corresponding to the conference, the distributor is responsible for executing the distribution logic.

[0108] 2. The assignor first identifies the new list of meeting numbers to be assigned by comparing the file names in the relevant directories.

[0109] 3. Read the load information of all servers in the group, and for a single meeting, assign a corresponding signaling server to each meeting number according to the allocation strategy of not overloading and with the lightest load, and write the result into the first temporary file of the node directory monitored by the corresponding signaling server.

[0110] 4. The signaling server reads the list of conference numbers for which it is responsible, and creates result files corresponding to the conference numbers in the conference directory one by one. The result files contain the subscriber information such as the creator's node id, ip, port, etc.

[0111] 5. Finally, the server directly accessed by the terminal is notified by the conference publisher by monitoring whether the result file exists, and reads the relevant information of the working signaling server and returns it to the requesting terminal.

[0112] The process of the terminal requesting the signaling server is shown in the sequence diagram as follows: Figure 3 As shown, Figure 3 A specific example diagram of a video conference communication control method in an embodiment of the present invention includes the following:

[0113] 1) The first server that accepts terminal access is the conference publisher, and the access request sent by the terminal carries the conference number;

[0114] 2) The publisher creates the / group / subjects / meeting number file, and the assignor always monitors the changes of the files under the meeting directory / group / subjects / (i.e., monitors the changes of the / group / subjects / subdirectory). When it is detected that there is only a meeting number file such as " / stp / subjects / 123456" for a certain meeting number, but no result file corresponding to the meeting number, it is used as the meeting number to be assigned;

[0115] 3) The allocator reads all files in the / group / subscribers / directory of the cluster (i.e. reads the / group / subscribers subdirectory information), reads all node information, and allocates subscriber servers for the conference numbers to be allocated based on the load information;

[0116] 4) directly updating the conference number to be allocated into the first temporary file corresponding to the subscriber server;

[0117] 5) Each subscriber continuously monitors the node directory, and after monitoring that its first temporary file has changed, it reads the newly assigned conference number to be assigned, and immediately creates a result file in the conference directory. The result file is named with the conference number to be assigned, for example, with the conference number to be assigned as a prefix;

[0118] 6) The publisher continuously monitors the changes in the conference directory. When a change is detected, the publisher obtains the result file of the conference number to be assigned, thereby obtaining the information of the corresponding subscriber and feeding it back to the terminal.

[0119] In the embodiment of the present invention, the Zookeeper cluster supports multiple video conference processing at the same time, refer to Figure 4 , Figure 4 As another specific example of the video conference communication control method in the embodiment of the present invention, Zookeeper provides a set of client external interfaces, which are as follows:

[0120] 1. Connect to the Zookeeper cluster;

[0121] 2. Create a directory;

[0122] Zookeeper supports the creation of three types of directories: temporary directories, persistent directories, and persistent directories with expiration dates. If the created directory is a temporary directory, the Zookeeper service will automatically delete all temporary directories created by the session when the session between the ZookeeperClient instance and the Zookeeper service ends normally or abnormally. Persistent directories must be explicitly deleted, otherwise they will always exist; persistent directories with expiration dates will be automatically deleted by the Zookeeper service when the expiration date expires, and the end of the creator's session will have no effect on them.

[0123] 3. Get directory data (supports "monitoring" semantics, that is, if the directory data changes, a notification is generated);

[0124] 4. Update catalog data;

[0125] 5. Delete the directory;

[0126] 6. Check whether the directory exists (supports "monitoring" semantics, that is, notification is generated when the directory is created or disappears);

[0127] 7. Get the sub-directory information of a directory (supports "monitoring" semantics, i.e., a notification is generated if there is any change in the sub-directory information of a directory).

[0128] The schematic diagram of the interaction between the signaling server and the terminal involved in the Zookeeper network is as follows: Figure 4 shown.

[0129] Figure 5 FIG. 2 is another specific example of the video conference communication control method in the embodiment of the present invention. Figure 5 As shown, in this method:

[0130] 1) Each signaling server that joins the group:

[0131] 1. Join the group: create a temporary directory: / group / subscribers / nodeId, the directory data includes the signaling server address and node ID, etc.; and monitor the changes of its directory information (the assignor will change);

[0132] 2. Update load: Create a temporary directory: / group / subscribers / nodeId_info. The directory data is the load status of the signaling server. If the load changes after joining the group, it will be updated;

[0133] 3. Competing allocator: Create a temporary directory (with node ID): / group / router / master; if the directory is created successfully, or if the directory already exists but the read directory information (node ​​ID) is the same as its own, it is considered successful. After becoming an allocator, monitor the changes of the subdirectories of the directories / group / subjects / and / group / subscribers / ; if it fails, monitor the changes of the directory / group / router / master itself;

[0134] 2) Publisher:

[0135] When the terminal carries the conference number to request the signaling server address, it creates a persistent directory with a validity period: / group / subjects / conference number, and monitors the directory / group / subjects / conference number_Owner, waiting for its appearance;

[0136] 3) Allocator:

[0137] Found the change of / group / subjects subdirectory, obtained the meeting number to be assigned by comparing the difference between the meeting number and the meeting number_owner in the directory, obtained the / group / subscribers / subdirectory information, obtained the communication and load information of all signaling servers in the group, determined the signaling server corresponding to the meeting number according to the load balancing strategy, and updated its corresponding file information: / group / subscribers / nodeId, and added the meeting number to the meeting list responsible for the node;

[0138] 4) Subscribers:

[0139] By monitoring the node directory data to find changes, read the list of pending meetings from the directory / group / subscribers / nodeId, create the corresponding / group / subjects / meeting number_Owner one by one, and write the communication information of its own node;

[0140] 5)Publisher:

[0141] It is found that / group / subjects / meeting number_owner is generated in the conference directory, and the communication information of the signaling server corresponding to the conference number is notified to the upper-level module and returned to the terminal.

[0142] It should be noted that Figure 5 Monitoring or discovering directories / subdirectories means monitoring or discovering data changes in large directories. Creating a directory also means creating files in the directory at the same time.

[0143] The following is a detailed description of how the method described in the embodiment of the present invention uses a set of interfaces provided by the Zookeeper client to achieve the uniqueness of the signaling server address allocated and synchronized to the terminal, in combination with three important business scenarios: (1) server node startup; (2) the terminal obtains the signaling server address corresponding to the meeting from any server; and (3) server node failure.

[0144] The following examples all have the same premise: the cluster group name is stp, the meeting number is 123456, the group has three servers, and the node IDs (illustration) are 31540f2c1_1, 31540f2c1_2, and 31540f2c1_3 respectively.

[0145] 1. The specific logic corresponding to the server node startup is as follows:

[0146] 1. Join a group

[0147] When the three servers are started, they create temporary directories and first temporary files: " / stp / subscribers / 31540f2c1_1", " / stp / subscribers / 31540f2c1_2", and " / stp / subscribers / 31540f2c1_3". The data in the first temporary file is the communication address and node ID published by the server, and the data changes in the first temporary file are monitored (the distributor will write the meeting number that the node is responsible for). The node ID prefix is ​​the newly generated guid, which can ensure global uniqueness.

[0148] 2. Update Load

[0149] When the three signaling servers are started, the second temporary files are created respectively: " / stp / subscribers / 31540f2c1_1_info", " / stp / subscribers / 31540f2c1_2_info", and " / stp / subscribers / 31540f2c1_3_info". The data in the second temporary files are the server load status, CPU, memory, number of connections and other information. If the node load changes after joining the group, it needs to be updated in real time.

[0150] 3. Competitive allocators

[0151] When the three servers start, they try to create the distributor directory: " / stp / router / master". If the creation is successful, they become distributors and monitor the changes of the conference directory " / stp / subjects / " and the node directory " / stp / subscribers / ". If the competition for distributor fails, they monitor the changes of the file " / stp / router / master" itself, which will be automatically deleted when the distributor fails. For example, if node 2 becomes the distributor, then " / stp / router / master" on the Zookeeper server is created by node 2, and its corresponding directory data is: 2#10.130.93.108:6000.

[0152] 2. When the terminal requests the signaling server address corresponding to the conference from server node 1, the specific logic of the server is as follows:

[0153] When server node 1 receives the request command (carrying the conference number 123456) from the terminal, it creates a persistent directory with a validity period (for example, 72 hours): " / stp / subjects / 123456", and monitors the file " / stp / subjects / 123456_Owner", waiting for its appearance.

[0154] Server node 2, which plays the role of allocator, discovers the file information and changes in the conference directory " / stp / subjects / " through monitoring, and finds that " / stp / subjects / 123456" exists, but the corresponding " / stp / subjects / 123456_Owner" does not exist, and obtains the conference number 123456 to be allocated. Then the allocator obtains all file information in " / stp / subscribers / ", obtains the communication and load information of all servers in the stp group, and decides on the signaling server corresponding to the conference number 123456 according to the allocation strategy of not overloading and having the lightest load. The allocation logic here can refer to Table 1 and Table 2. Table 1 divides the node load into four gears, 1 / 2 / 3 / overload. The lower the gear, the lighter the load. Assuming the current load conditions of the three signaling servers are as shown in Table 2, according to the strategy of not overloading and having the lightest load, the working signaling server corresponding to the conference number 123456 is node 3. For specific allocation logic, see Table 1 and Table 2. Figure 6 .

[0155] Table 1

[0156] Gear CPU usage Memory usage Number of connections 1 0~30% 0~30% -- 2 30%~60% 30%~60% -- 3 60%~80% 60%~80% -- Overload 80%~100% 80%~100% --

[0157] Table 2

[0158] Node Directory CPU usage Memory usage Number of connections / stp / subscribers / 31540f2c1_1 90% 10% 1 / stp / subscribers / 31540f2c1_2 25% 50% 300 / stp / subscribers / 31540f2c1_3 20% 40% 200

[0159] Figure 6 FIG. 2 is another specific example of the video conference communication control method in the embodiment of the present invention. Figure 6 As shown, using the load information to determine the subscriber corresponding to the conference identifier to be allocated may include:

[0160] 1. Exclude all overloaded nodes in the candidate list. If any of the CPU / memory items is overloaded, it is considered overloaded.

[0161] 2. Select the node with the lowest CPU level in the candidate list;

[0162] 3. Whether there is more than one node with the lowest CPU level;

[0163] 4. If yes, continue to select the node with the lowest archive position in the node list with the lowest CPU gear; then determine whether there is more than one node with the lowest archive position. If yes, continue to select the node with the least number of connections in the node list with the lowest archive position to obtain the appropriate node;

[0164] 5. If not, get the appropriate node.

[0165] Subsequently, the distributor will update the meeting number 123456 to the corresponding file in the node directory: " / stp / subscribers / 31540f2c1_3", that is, add the meeting number 123456 to the list of meetings that the node is responsible for.

[0166] Subscriber node 3 discovers changes by monitoring the node directory data, reads the pending conference number 123456 from its corresponding first temporary file " / stp / subscribers / 31540f2c1_3", and creates the corresponding " / stp / subjects / 123456_Owner" file in the conference directory. The data content passed in when creating the file is: the communication information of node 3. Due to various reasons, when multiple nodes are concurrently created to create the file " / stp / subjects / 123456_Owner", the Zookeeper service can ensure the atomicity of the creation interface, either success or failure, and return a specific error code if it fails, thus ensuring the uniqueness of the result file.

[0167] The server node 1 directly accessed by the terminal finds the generation of " / stp / subjects / 123456_Owner" through monitoring, calls the interface for obtaining directory data, obtains the signaling server information finally assigned to the conference as node 3, calls back to the upper layer, and returns it to the terminal after being processed by the upper layer module.

[0168] 3. Business scenarios of server failure

[0169] (1) The server is a distributor but is not responsible for any conferences and is not a subscriber to any conferences.

[0170] After the allocator server fails, the session between the ZookeeperClient loaded by it and the Zookeeper service ends abnormally, and the temporary directory / group / router / master created by it will be automatically deleted by the Zookeeper service. When other servers join the group, they monitor the changes of the / group / router / master file through the ZookeeperClient interface "check whether the directory exists" (with monitoring semantics), and can be notified, triggering a new round of "allocator competition" logic. Multiple servers concurrently create the directory / group / router / master. The Zookeeper service ensures the atomicity of the interface, so only one server can create it successfully, and the other servers will fail to create, and start a new round of monitoring of the directory / group / router / master.

[0171] Before the allocator server fails, the last possible states are as follows:

[0172] 1. The meeting number to be allocated is compared, but the allocation logic has not yet been executed;

[0173] 2. The allocation logic for the conference number has been executed, but the result has not yet been synchronized to the Zookeeper service;

[0174] 3. The directory corresponding to the allocated conference subscriber server: / group / subscribers / nodeId has been synchronized to the new conference number through the Zookeeper service, but has not yet been created: / group / subjects / conference number_owner;

[0175] 4. The allocated conference subscriber server successfully creates / group / subjects / conference number_owner for the new conference number.

[0176] The logic of the new allocator that succeeds in the competition needs to be compatible with the above four situations. The logic of the new allocator is as follows:

[0177] Actively obtain the file information of the conference directory / group / subjects, compare the difference between the conference number and the conference number_owner in the directory, get the conference number to be assigned, obtain the second temporary file information of the server in the node directory / group / subscribers / , obtain the communication and load information of all servers in the group, and determine the signaling server corresponding to the conference number according to the allocation strategy of not overloading and with the lightest load, and update its corresponding file: / group / subscribers / nodeId, and add the conference number to the conference list responsible for the subscriber node.

[0178] For situations 1 and 2, the logic of the new allocator will not miss the conference number to be allocated that is in a critical state; for situation 4, when the logic of the new allocator compares the directory, if the conference number already has a corresponding "conference number__owner", then the conference number will not generate new allocation logic; for situation 3, due to the difference in network communication quality between nodes, it is possible that the new allocator generates a new signaling server after executing the allocation logic, and then the new signaling server corresponding to the conference number and the signaling server in the synchronization state allocated by the old allocator will simultaneously create the file: / group / subjects / conference number_owner. In this case, the atomicity of the Zookeeper service to create the directory interface is relied on to ensure the uniqueness of the result. There will only be one signaling server that successfully creates the file: / group / subjects / conference number_owner, and other concurrent signaling servers fail to create and obtain a clear error code. In the embodiment of the present invention, error codes will be given for different error reports, and different error codes have corresponding explanations. An error code description file can be preset, and the error code description file includes the error code and the fault analysis corresponding to the error code, which is convenient for quick query and problem solving.

[0179] Another situation that needs to be mentioned is that if the allocator server failure is not a machine or process failure, but just a short-term poor network communication, then after the network is restored, the logic of the old allocator must also not affect the uniqueness of the result. The session keepalive tolerance timeout set by the client is usually shorter and more sensitive than that of the server. When the Zookeeper service perceives the abnormality of the allocator session, the old allocator has already perceived the abnormal event of the session disconnection. After the abnormality occurs, the session is interrupted. After the network is restored again, the old allocator needs to first rejoin the server group through the interface "Connect to Zookeeper Cluster". After joining successfully, execute the logic of "Competing Allocators". Generally speaking, in most cases, after the network is restored, since the new allocator has been generated, the old allocator will fail to create the file / group / router / master, and will get the error code "File already exists", and start monitoring the changes of the file / group / router / master.

[0180] (2) The server is responsible for the conference and is a subscriber but not a distributor.

[0181] After the server fails, the session between the ZookeeperClient loaded by it and the Zookeeper service ends abnormally. Then the temporary file / group / subscribers / NodeId and the file / group / subjects / NodeId_info created by it, as well as the file / group / subjects / ConferenceNumber_owner, will be automatically deleted by the Zookeeper service. Since the assignor monitors the / group / subscribers / subdirectory, it will sense the change and re-execute the logic of comparing the unassigned conference numbers, assigning signaling servers, and creating the corresponding file: / group / subjects / ConferenceNumber_owner for the assigned new signaling server. After the terminal senses the abnormality of the signaling server used through the heartbeat timeout, when it requests a new signaling server address again, it can directly read the data of the newly created file: / group / subjects / ConferenceNumber_owner, obtain the newly assigned signaling server information, and return it to the terminal.

[0182] (3) The server is both a distributor and a subscriber.

[0183] After the server fails, the directory / group / router / master and file / group / subjects / Conference Number_owner created by it will be automatically deleted by the Zookeeper service. Other servers perceive the change in the directory / group / router / master, triggering the execution of the "competitive allocator" logic to generate a new allocator. When the new allocator is generated, it will compare the unassigned conference numbers, assign signaling servers, and create the corresponding file: / group / subjects / Conference Number_owner for the assigned new signaling server. When the terminal perceives the abnormality of the signaling server used through the heartbeat timeout, when requesting a new signaling server address again, it can directly read the data of the newly created file: / group / subjects / Conference Number_owner, obtain the newly assigned signaling server information, and return it to the terminal.

[0184] In summary, the embodiment of the present invention provides a method for implementing uniqueness of signaling node allocation by using Zookeeper networking in a video conference, which mainly has the following features:

[0185] (1) The master node is elected to allocate conference services. The master node makes election decisions through Zookeeper.

[0186] (2) The master node perceives the load information of each node through Zookeeper, and the allocation algorithm uses load balancing scheduling;

[0187] (3) The node notifies and confirms tasks with other nodes through Zookeeper;

[0188] (4) The master node tracks the survival status of each node through Zookeeper; when a node failure is found, the node tasks are recycled and load balanced;

[0189] (5) The consistency of intermediate states is achieved through the characteristics of Zookeeper.

[0190] According to the method of the embodiment of the present invention, the allocation is performed. Even in the scenario where there are situations such as node birth and death, terminal requests concurrently, or even poor network communication, the uniqueness of the signaling server allocated to the same conference in the video conference can still be accurately guaranteed, and the routing allocation between different conferences can be load balanced as much as possible. When any server fails, it can automatically identify the abnormality and promptly allocate a new server with a unique corresponding conference number to the conference list undertaken by the failed server. At the same time, the embodiment of the present invention can accurately identify faults. This method is not only applicable to existing conferences in the conference system, but also to newly created conferences through various business means.

[0191] The present invention also provides a video conference communication control device, as described in the following embodiments. Since the principle of the device to solve the problem is similar to that of the video conference communication control method, the implementation of the device can refer to the implementation of the video conference communication control method, and the repeated parts will not be repeated.

[0192] Figure 7 FIG. 1 is a schematic diagram of a video conference communication control device according to an embodiment of the present invention. Figure 7 As shown, the device comprises:

[0193] The startup processing module 701 is used to create a node directory when the server is started; the node directory is used to update the storage server load information and temporarily store the conference identifier that the server is responsible for;

[0194] The distributor competition module 702 is used to compete for distributors; during the distributor competition process, a distributor directory is created, and distributor information is stored in the distributor directory; the distributor directory is automatically deleted when the distributor fails;

[0195] The distributor processing module 703 is used to monitor whether the conference identifier to be allocated appears in the conference directory when the competing distributor succeeds; the conference directory stores the conference identifier and the result file corresponding to the conference identifier. The conference identifier to be allocated is the conference identifier that has no corresponding result file in the conference directory. The conference directory is created by the publisher, and the publisher is the first server accessed by the terminal. The terminal access carries the conference identifier; the publisher is used to: create the conference directory when the terminal accesses for the first time; receive the terminal access, and provide the terminal with the subscriber information corresponding to the conference identifier by monitoring the changes in the conference directory; when the conference identifier to be allocated appears, the distributor obtains the load information of all servers in the cluster, and uses the load information to determine the corresponding conference identifier to be allocated. The subscriber updates the conference identifier to be allocated to the node directory, so that the subscriber corresponding to the conference identifier to be allocated: obtains the conference identifier to be allocated by monitoring the changes in the node directory, creates a result file according to the conference identifier to be allocated to the conference directory, and deletes the conference identifier to be allocated in the node directory; wherein the result file carries the conference identifier; wherein, when the server becomes any role, the role's monitoring function for the designated directory is started; the designated directory includes the node directory, the conference directory, and the distributor directory; servers with different roles monitor different designated directories; when the publisher and subscriber monitor and obtain the message that the distributor directory is deleted, they start to compete for the distributor; when the subscriber fails, the failed result file in the conference directory is automatically deleted.

[0196] In one embodiment, the same server plays the role of one or any combination of publisher, distributor, subscriber.

[0197] In one embodiment, the distributed coordination service includes an atomic broadcast coordination function.

[0198] In one embodiment, the distributor processing module is specifically used to:

[0199] Based on the atomic broadcast coordination function, try to create the distributor directory; the atomic broadcast coordination function is used to ensure that there is only one server in the cluster that can successfully create the distributor directory;

[0200] When the distributor directory is created successfully, you become the distributor.

[0201] In one embodiment, the conference identifier stored in the conference directory exists in the form of a file named with the conference identifier, and the result file stored in the conference directory exists in the form of a file named with the conference identifier and the first character identifier; the first character is arbitrarily specified;

[0202] The allocator processing module is specifically used to:

[0203] By comparing the file names, the conference ID without result files is determined, and the conference ID without result files is used as the conference ID to be allocated.

[0204] In one embodiment, the subscriber creates a node directory, which is used to store the subscriber's basic server information, load information, and conference identifiers responsible for processing conferences.

[0205] In one embodiment, the node directory includes multiple pairs of temporary files, each pair of temporary files includes a first temporary file and a second temporary file, the first temporary file updates and stores basic information of the server in real time, temporarily stores the conference identifier responsible for processing the conference, and the second temporary file updates and stores load information of the same server as in the first temporary file in real time;

[0206] When a subscriber fails, the first temporary file and the second temporary file corresponding to the failed subscriber are automatically deleted.

[0207] In one embodiment, the load information includes CPU usage, memory usage, and number of connections;

[0208] The allocator processing module is specifically used to:

[0209] The CPU usage, memory usage, and number of connections of each server in the cluster are used to determine the subscriber corresponding to the conference ID to be allocated.

[0210] An embodiment of the present invention further provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned video conference communication control method when executing the computer program.

[0211] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned video conference communication control method is implemented.

[0212] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the above-mentioned video conference communication control method is implemented.

[0213] In an embodiment of the present invention, for any server in a distributed cluster, when the server is started, a node directory is created; the node directory is used to: update the storage server load information, temporarily store the conference identifier that the server is responsible for; compete for the allocator; create an allocator directory in the process of competing for the allocator, and store the allocator information in the allocator directory; the allocator directory is automatically deleted when the allocator fails; when the competing allocator succeeds: monitor whether the conference identifier to be allocated appears in the conference directory; the conference directory stores the conference identifier and the result file corresponding to the conference identifier, the conference identifier to be allocated is the conference identifier that does not have a corresponding result file in the conference directory, the conference directory is created by the publisher, the publisher is the first server accessed by the terminal, and the terminal accesses the conference identifier; the publisher is used to: create the conference directory when the terminal accesses for the first time; receive the terminal access, and provide the terminal with the corresponding information of the conference identifier by monitoring the changes in the conference directory Subscriber information; when a conference identifier to be allocated appears, the distributor obtains the load information of all servers in the cluster, and uses the load information to determine the subscriber corresponding to the conference identifier to be allocated, and updates the conference identifier to be allocated to the node directory, so that the subscriber corresponding to the conference identifier to be allocated: obtains the conference identifier to be allocated by monitoring the changes in the node directory, creates a result file in the conference directory according to the conference identifier to be allocated, and deletes the conference identifier to be allocated in the node directory; wherein the result file carries the conference identifier; wherein, when the server becomes any role, the role's monitoring function for the specified directory is started; the specified directory includes the node directory, the conference directory, and the distributor directory; servers with different roles monitor different specified directories; when the publisher or subscriber monitors and obtains the message that the distributor directory is deleted, it starts to compete for the distributor; when the subscriber fails, the failed result file in the conference directory is automatically deleted. By using the video conference communication control method in the embodiment of the present invention to allocate working signaling servers, in various scenarios where subscriber failures, distributor failures, publisher failures, terminal request concurrency, and even poor network communication occur, the uniqueness of the working signaling server allocated to the same conference in the video conference can still be accurately guaranteed, and the routing allocation between different conferences can achieve load balancing as much as possible. When the working signaling server fails, it can automatically identify the anomaly and promptly allocate a suitable and unique new server to the conference undertaken by the failed server. The video conference communication control method is not only applicable to existing conferences in the conference system, but also to newly created conferences through various business means.

[0214] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0215] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0216] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0217] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0218] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A video conference communication control method, It is characterized in that Applied to any server in a distributed cluster, the server has the following role types: publisher, distributor, subscriber; The distributed cluster has a distributed coordination service, the distributed coordination service includes multiple components that provide consistency, synchronization and coordination functions for the distributed cluster, and the method includes: When the server starts, a node directory is created; the node directory is used to update the storage server load information and temporarily store the conference ID that the server is responsible for; Competing allocators; creating an allocator directory during the competing allocator process, storing allocator information in the allocator directory; the allocator directory is automatically deleted when the allocator fails; When a competing allocator succeeds: Monitor whether the conference directory appears with the conference ID to be allocated; the conference directory stores the conference ID and the result file corresponding to the conference ID. The conference ID to be allocated is the conference ID that does not have a corresponding result file in the conference directory. The conference directory is created by the publisher, which is the first server accessed by the terminal. The terminal accesses the conference with the conference ID; the publisher is used to: create the conference directory when the terminal accesses for the first time; receive terminal access, and provide the terminal with the subscriber information corresponding to the conference ID by monitoring the changes in the conference directory; When a conference identifier to be allocated appears, the load information of all servers in the cluster is obtained, and the subscriber corresponding to the conference identifier to be allocated is determined by using the load information, and the conference identifier to be allocated is updated to the node directory, so that the subscriber corresponding to the conference identifier to be allocated: the conference identifier to be allocated is obtained by monitoring the changes in the node directory, and a result file is created according to the conference identifier to be allocated to the conference directory, and the conference identifier to be allocated in the node directory is deleted at the same time; wherein the result file carries the conference identifier; Among them, when the server becomes any role, the role's monitoring function for the designated directory is started; the designated directory includes the node directory, the conference directory, and the distributor directory; different role servers monitor different designated directories; when the publisher and subscriber monitor and receive the message that the distributor directory is deleted, they start to compete for the distributor; when the subscriber fails, the result file of the failure in the conference directory is automatically deleted.

2. The method according to claim 1, It is characterized in that The same server can play the role of publisher, distributor, subscriber, or any combination thereof.

3. The method according to claim 1, It is characterized in that The distributed coordination service includes an atomic broadcast coordination function.

4. The method according to claim 3, It is characterized in that Competing allocators include: Based on the atomic broadcast coordination function, try to create the distributor directory; the atomic broadcast coordination function is used to ensure that there is only one server in the cluster that can successfully create the distributor directory; When the distributor directory is created successfully, you become the distributor.

5. The method according to claim 1, It is characterized in that The conference ID stored in the conference directory exists in the form of a file named with the conference ID, and the result file stored in the conference directory exists in the form of a file named with the conference ID and the first character ID; the first character is arbitrarily specified; Monitor the conference directory for the presence of conference IDs to be assigned, including: By comparing the file names, the conference ID without result files is determined, and the conference ID without result files is used as the conference ID to be allocated.

6. The method according to claim 1, It is characterized in that The node directory includes multiple pairs of temporary files, each pair of temporary files includes a first temporary file and a second temporary file, the first temporary file updates and stores basic information of the server in real time, temporarily stores the conference identifier responsible for processing the conference, and the second temporary file updates and stores load information of the same server as in the first temporary file in real time; When a subscriber fails, the first temporary file and the second temporary file corresponding to the failed subscriber are automatically deleted.

7. The method according to claim 1, It is characterized in that The load information includes CPU usage, memory usage, and number of connections; Using the load information, determining a subscriber corresponding to the conference identifier to be allocated includes: The CPU usage, memory usage, and number of connections of each server in the cluster are used to determine the subscriber corresponding to the conference ID to be allocated.

8. A video conference communication control device, It is characterized in that Applied to any server in a distributed cluster, the server has the following role types: publisher, distributor, subscriber; The distributed cluster has a distributed coordination service, and the distributed coordination service includes multiple components that provide consistency, synchronization and coordination functions for the distributed cluster. The device includes: The startup processing module is used to create a node directory when the server starts; the node directory is used to update the storage server load information and temporarily store the conference identifier that the server is responsible for; The distributor competition module is used to compete for distributors; during the distributor competition process, a distributor directory is created, and distributor information is stored in the distributor directory; the distributor directory is automatically deleted when a distributor fails; The distributor processing module is used to monitor whether the conference identifier to be allocated appears in the conference directory when the competing distributor succeeds; the conference directory stores the conference identifier and the result file corresponding to the conference identifier. The conference identifier to be allocated is the conference identifier that has no corresponding result file in the conference directory. The conference directory is created by the publisher, and the publisher is the first server accessed by the terminal. The terminal access carries the conference identifier; the publisher is used to: create the conference directory when the terminal accesses for the first time; receive terminal access, and provide the terminal with the subscriber information corresponding to the conference identifier by monitoring the changes in the conference directory; when the conference identifier to be allocated appears, the distributor obtains the load information of all servers in the cluster, and uses the load information to determine the subscriber information corresponding to the conference identifier to be allocated. Reader, update the conference identifier to be allocated to the node directory, so that the subscriber corresponding to the conference identifier to be allocated: obtain the conference identifier to be allocated by monitoring the changes in the node directory, create a result file according to the conference identifier to be allocated to the conference directory, and delete the conference identifier to be allocated in the node directory; wherein the result file carries the conference identifier; wherein, when the server becomes any role, the role's monitoring function for the specified directory is started; the specified directory includes the node directory, the conference directory, and the distributor directory; different role servers monitor different specified directories; when the publisher and subscriber monitor and obtain the message that the distributor directory is deleted, they start to compete for the distributor; when the subscriber fails, the failed result file in the conference directory is automatically deleted.

9. The device as claimed in claim 8, It is characterized in that The same server can play the role of publisher, distributor, subscriber, or any combination thereof.

10. The device according to claim 8, It is characterized in that The distributed coordination service includes an atomic broadcast coordination function.

11. The device according to claim 10, It is characterized in that The allocator processing module is specifically used to: Based on the atomic broadcast coordination function, try to create the distributor directory; the atomic broadcast coordination function is used to ensure that there is only one server in the cluster that can successfully create the distributor directory; When the distributor directory is created successfully, you become the distributor.

12. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

13. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

14. A computer program product, It is characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.