Audio and video cluster management method capable of dynamically expanding capacity

Abstract

The invention provides an audio and video cluster management method capable of dynamically expanding capacity, and belongs to the technical field of audio and video cluster management. Wherein all cluster nodes and the unified interface service are registered in the cluster service according to node types, the registered nodes send heartbeat and node resource occupation information to the management module at regular time, and the cluster management module realizes load balancing based on node report information to guarantee normal forwarding of streaming media data. If cluster resources are insufficient, a user can create a new node, the node is registered in the cluster service through the interface service, after registration is completed, a cluster management module monitors heartbeat and resource occupation information of the new node, subsequent tasks can be distributed to the new node based on a scheduling strategy and a node load state, and the new node will establish an inter-node data transmission channel; and completing a data transmission task between the nodes. The method has the characteristic of flexible expansion, and the stability of cluster operation is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of audio and video cluster management, and in particular relates to a dynamically expandable audio and video cluster management method. Background Art

[0002] With the widespread application of audio and video communication services in various fields, single-node servers are subject to their own resource limitations when processing large-scale audio and video services, and are prone to problems such as excessive CPU load and insufficient bandwidth, resulting in failure of media data forwarding and inability to maintain normal communication services.

[0003] At present, when resources are insufficient, audio and video services deployed in clusters need to stop services for expansion, resulting in the inability to restore services in a short period of time, and restarting services will cause various unstable factors. Summary of the invention

[0004] In view of this, the purpose of the present invention is to overcome the shortcomings of the prior art and provide a dynamically expandable audio and video cluster management method. The method dynamically adds audio and video service nodes during cluster operation, greatly improving the flexibility and scalability of the cluster, and can solve the dynamic expansion problem during the operation of audio and video services without affecting the normal operation of audio and video services.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] A dynamically scalable audio and video cluster management method comprises the following steps:

[0007] Step 1: The cluster starts normally, automatically confirms the identity, and starts node registration, node shutdown, status reporting, and node scheduling services based on the RPC protocol;

[0008] Step 2: Register the unified interface service;

[0009] Step 3: When there is no registered node in the cluster or all nodes in the cluster are currently running at full load, manually add and start the audio and video node. The audio and video node reads the configuration file, obtains the address of the cluster management module, and sends a registration request to the cluster management module using the node type, node ID, and node load parameters as request parameters.

[0010] Step 4: After receiving the audio and video node registration request, the cluster management module stores the request parameter information. If the registration is successful, a success message is returned. If the registration fails and the current registration times are less than the preset upper limit, re-registration is performed. If the registration is successful or the registration times are higher than the preset upper limit, the registration is stopped.

[0011] Step 5: The cluster management module monitors the heartbeat of the registered nodes once per second to ensure that the audio and video nodes are active. If the heartbeat of the node cannot be detected for multiple consecutive times, the node will be deleted from the cluster. The deleted node needs to be re-registered after it returns to normal.

[0012] Step 6: The successfully registered audio and video nodes report the current load information every second;

[0013] Step 7: After receiving the current load information reported by the node, the cluster management module compares it with the load threshold of the node obtained by the cluster management module during the node registration phase. If the current load is within the load threshold, the node can undertake the task. If the current load exceeds the load threshold, no task will be assigned to the node.

[0014] Step 8: After receiving a new service request, the cluster management module allocates the task to the registered nodes according to the task type in the request and the load balancing strategy;

[0015] Step 9: After the node receives the task, it can be divided into the following two situations:

[0016] 1) If a new data consumption task is assigned to the previously registered node A, node A establishes a connection with the new consumer, so that the media data is transferred from the original producer to the new consumer;

[0017] 2) If a data consumption task is assigned to the newly registered node B, node B will establish a room to connect with the new consumer, establish a data transmission channel between nodes, and initiate a data channel establishment request to node A;

[0018] After receiving the data channel establishment request from node B, node A also establishes the data transmission channel between nodes and sends a data channel connection request to node B. At this time, node A has a data producer, and the connection between the data producer and the data transmission channel will be established at the same time;

[0019] After node B receives the data channel connection request from node A, it establishes a connection with node A. At this time, the complete data channel has been established, and the new data consumer in node B can consume the audio and video data generated by the producer in node A, realizing dynamic expansion.

[0020] Furthermore, in step 6, if the audio and video node is limited by the CPU load, the CPU usage rate is reported; if the audio and video node is limited by the remaining disk space, the remaining disk space is reported.

[0021] The beneficial effects of the present invention are:

[0022] 1. The present invention is based on message queues and RPC protocols, integrates resource monitoring and exception handling mechanisms, supports dynamic expansion of multiple types of audio and video service nodes during cluster service operation, and effectively enhances cluster operation stability.

[0023] 2. The present invention improves the availability or flexible scalability of the audio and video cluster and reduces the operation and maintenance cost of the audio and video cluster. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a flow chart of an embodiment of the present invention.

[0025] Figure 2 The figure is a schematic diagram of the principle of constructing a data transmission channel between nodes in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The present invention is described in detail below with reference to the accompanying drawings.

[0027] A dynamically scalable audio and video cluster management method, in which all cluster nodes and unified interface services are registered in the cluster service according to the node type. The registered nodes periodically send heartbeats and node resource occupancy information to the management module. The cluster management module implements load balancing based on the node reporting information to ensure the normal forwarding of streaming media data. If the cluster resources are insufficient, the user can create a new node, which is registered to the cluster service via the interface service. After the registration is completed, the cluster management module monitors the heartbeat and resource occupancy (load) information of the new node. Subsequent tasks can be assigned to the new node based on the scheduling strategy and node load status. The new node will establish an inter-node data transmission channel to complete the inter-node data transmission task.

[0028] Specifically, if Figure 1 As shown, the method comprises the following steps:

[0029] (1) After the cluster is started, the identity is automatically confirmed, and node registration, node shutdown, status reporting, and node scheduling services are enabled based on the RPC protocol;

[0030] (2) Register for the unified interface service;

[0031] (3) When there are no registered nodes in the cluster or all nodes in the cluster are currently running at full capacity, dynamic capacity expansion is required. Manually add and start the audio and video node. The audio and video node reads the configuration file, obtains the address of the cluster management module, and sends a registration request to the cluster management module using the node type (service type), node ID, and node load parameters as request parameters. Figure 1 As shown;

[0032] (4) The cluster management module receives the audio and video node registration request and stores the node request with parameter information. If the registration is successful, a success message is returned. If the registration fails and the current number of registrations is less than the set upper limit, the registration is re-registered. If the registration is successful or the number of registrations is higher than the set upper limit, the registration is stopped. Figure 1 As shown;

[0033] (5) The cluster management module monitors the heartbeat of registered nodes once per second to ensure that the audio and video nodes are active. If the node cannot be detected three times in a row, the node will be deleted from the cluster. The deleted node can be re-registered after it returns to normal.

[0034] (6) The audio and video nodes that have been successfully registered report the current load information every second. The load information type can be reported according to the type of audio and video node. Taking CPU usage as an example, the node sets the maximum CPU usage to 0.7. If the current reported CPU usage is 0.3, the node can still undertake tasks. If the current reported CPU usage is 0.75, no tasks will be assigned to the node.

[0035] (7) After receiving the current load information reported by the node, the cluster management module compares it with the load threshold of the node obtained by the cluster management module during the node registration phase. If the current load is within the load threshold, the node can undertake the task. If the current load exceeds the load threshold, no task will be assigned to the node.

[0036] (8) After receiving the service request, the cluster management module will match the nodes with the service type of "video conferencing" from the registered audio and video nodes according to the task type in the request, such as "video conferencing", and allocate the task to the subordinate nodes according to the strategy and load conditions;

[0037] (9) After a node receives a task, it can be divided into the following two situations:

[0038] 1) If there is only one node A, it means that all tasks are running in one audio and video node. The new data consumption task is assigned to the previously registered node A. Node A will establish a connection with the new consumer, so that media data can be transferred from the original producer to the new consumer.

[0039] 2) If node B executes the same task as node A, that is, there are currently two audio and video nodes, then Figure 2 As shown:

[0040] Node B establishes a data transmission channel and sends a request to other nodes (such as node A) to establish an inter-node data transmission channel. Figure 2Identifier 1 in; if there is a data consumer in node B, after node B establishes the inter-node transmission channel, it connects the node B data consumer with the data transmission channel in this node, such as Figure 2 The logo 2 in;

[0041] The audio and video node A receives the request to establish an inter-node data transmission channel and establishes an inter-node data transmission channel. Figure 2 After the identification 3 in the figure is successfully established, a node-to-node data channel connection request is sent to node B. If there is a data producer in node A, the data producer in node A is connected to the node-to-node data transmission channel after the node-to-node data transmission channel is established. Figure 2 The logo 4 in;

[0042] Node B receives the connection request from node A and connects to the data transmission channel between node A. Figure 2 Middle logo 5;

[0043] If there is a data producer in node A, a channel data consumer is created in this node. After the creation is successful, a request to create a channel data producer is sent to node B.

[0044] Node B receives the request from node A, creates a channel data producer, and completes the data transmission process from node A to node B.

[0045] In summary, the present invention integrates cluster resource monitoring and exception handling mechanisms, and supports dynamic expansion of different types of audio and video nodes in audio and video clusters during service operation through message queues and RPC protocols. The present invention does not require resetting cluster services, and can achieve dynamic expansion of multiple types of service nodes in large-scale audio and video clusters. It has the characteristics of flexible expansion and improves the stability of cluster operation.

[0046] It should be noted that any modifications, equivalent substitutions, improvements, etc. made outside the spirit and principles of the present invention are included in the protection scope of the present invention.

Claims

1. A dynamically scalable audio and video cluster management method, characterized in that: The following steps are involved: Step 1: The cluster starts normally, automatically confirms the identity, and starts node registration, node shutdown, status reporting, and node scheduling services based on the RPC protocol; Step 2: Register the unified interface service; Step 3: When there is no registered node in the cluster or all nodes in the cluster are currently running at full load, manually add and start the audio and video node. The audio and video node reads the configuration file, obtains the address of the cluster management module, and sends a registration request to the cluster management module using the node type, node ID, and node load parameters as request parameters. Step 4: After receiving the audio and video node registration request, the cluster management module stores the request parameter information. If the registration is successful, a success message is returned. If the registration fails and the current registration times are less than the preset upper limit, re-registration is performed. If the registration is successful or the registration times are higher than the preset upper limit, the registration is stopped. Step 5: The cluster management module monitors the heartbeat of the registered nodes once per second to ensure that the audio and video nodes are active. If the heartbeat of the node cannot be detected for multiple consecutive times, the node will be deleted from the cluster. The deleted node needs to be re-registered after it returns to normal. Step 6: The successfully registered audio and video nodes report the current load information every second; Step 7: After receiving the current load information reported by the node, the cluster management module compares it with the load threshold of the node obtained by the cluster management module during the node registration phase. If the current load is within the load threshold, the node can undertake the task. If the current load exceeds the load threshold, no task will be assigned to the node. Step 8: After receiving a new service request, the cluster management module allocates the task to the registered nodes according to the task type in the request and the load balancing strategy; Step 9: After the node receives the task, it can be divided into the following two situations: 1) If a new data consumption task is assigned to the previously registered node A, node A establishes a connection with the new consumer, so that the media data is transferred from the original producer to the new consumer; 2) If a data consumption task is assigned to the newly registered node B, node B will establish a room to connect with the new consumer, establish a data transmission channel between nodes, and initiate a data channel establishment request to node A; After receiving the data channel establishment request from node B, node A also establishes the data transmission channel between nodes and sends a data channel connection request to node B. At this time, node A has a data producer, and the connection between the data producer and the data transmission channel will be established at the same time; After node B receives the data channel connection request from node A, it establishes a connection with node A. At this time, the complete data channel has been established, and the new data consumer in node B can consume the audio and video data generated by the producer in node A, realizing dynamic expansion.

2. A dynamically scalable audio and video cluster management method according to claim 1, characterized in that: In step 6, if the audio and video node is limited by the CPU load, the CPU usage rate is reported; if the audio and video node is limited by the remaining disk space, the remaining disk space is reported.

Images