TWI713357B - Video streaming using hierarchical data center failover method - Google Patents

Abstract

A failover method using hierarchical data center for audio and video streaming is characterized in that: the audio and video streaming provider uploads the streaming content to the data center network, and a main server backs up the content from top to bottom in a hierarchical structure At each peer of the data center network, a maintenance server confirms the location and number of clients in each region and configures the required redundant edge servers. When the client encounters network failures and other conditions that cause the streaming to be interrupted and discontinuous, the data center network uses the DNS mechanism to direct the client upwards to the nearest available backup server or P2P download point in a hierarchical structure. Since the transfer client connects to the backup server quite quickly, the user can connect to the backup server or P2P download point in other regions without knowing it and continue to watch the streaming content.

Description

Video streaming using hierarchical data center failover method

The present invention relates to a failover method for a hierarchical data center for audio and video streaming, wherein the hierarchical data center provides client connections to ensure good viewing quality, and at the same time backs up streaming content to backup servers in various regions Or P2P download point. When encountering server failure or network congestion and other situations that cause the streaming to be interrupted and discontinuous, the data center network uses the DNS mechanism according to the hierarchical structure to transfer the client first to the nearby backup server with better quality or P2P download At the same time, troubleshooting problems to ensure that the client has a good viewing quality; further applying the aforementioned technology, each data center will also direct the client to a higher-level data center through the DNS mechanism when the backup server is insufficient; and the present invention It also provides an implementation method for maintaining server configuration backup, which can maintain user viewing quality and improve user satisfaction.

Most of the online audio-visual industry nowadays choose to use Content Delivery Network (CDN) to provide audio-visual streaming services. CDN is a data center system connected to each other through the Internet. It uses the server closest to each user to deliver music, pictures, videos, applications or even live program content to users faster and more reliably.

The aforementioned basic CDN architecture requires operators to rent and purchase computer rooms and deploy servers to provide users with connection and use, but this construction method results in high equipment costs for operators. In some recent studies, researchers have combined P2P (point-to-point transmission technology) with CDN deployment and management technology to reduce business costs. this. Users participating in CDN deployment use their own computer hardware as the CDN's cache node, and they can also get bonuses, goods, or network services as a reward.

The challenge of applying CDN-P2P to audio and video streaming is that live broadcasts usually have real-time and continuity. For example, users need to watch continuous content before and after live news to understand the information in the report, but network congestion and network Delays or server failures on the provider side may cause discontinuities in live broadcast content, which will greatly reduce the impression of viewers on the industry. Sometimes the viewers are impatient and may even switch to the channels of other video and audio streaming companies, which leads to the loss of viewers.

Most of the current audio and video streaming technologies transfer related users to adjacent edge servers in the event of a failure, but none of the aforementioned technologies provide a specific failover mechanism with related server configuration and maintenance algorithms to ensure audio and video streaming Streaming service quality.

The purpose of the present invention is to provide a failover method for audio and video streaming using a hierarchical data center. First, the streaming content is distributed to backup servers in various regions in a hierarchical tree structure from top to bottom. When the client encounters network congestion, network delay, or server failure on the provider side, resulting in discontinuous streaming interruption, through the failover mechanism of the present invention, the relevant client preferentially connects to nearby devices with better network quality. Support server or P2P download point to ensure the quality of audio and video streaming services on the client side.

For this reason, the technology adopted in the present invention is as follows: any data center in this case is structured on a hierarchical data center, basically including an audio-visual streaming provider, a data center network, multiple P2P networks, and multiple users end.

The foregoing audiovisual streaming provider provides and uploads audiovisual streaming content to the foregoing data center network.

The aforementioned data center network includes a main server, a maintenance server, at least one DNS server, and multiple backup servers. The aforementioned main server stores the video and audio streaming content from the provider, and sends the streaming content to the backup server where each client is located. The backup server stores the backup of the video and audio streaming content from the main server. When the client encounters network congestion, network delay, or failure of the provider server, which causes the streaming to be interrupted and discontinuous, the backup server will directly provide a backup of the streaming content for the client to watch.

The aforementioned P2P network is composed of multiple client terminals as a delivery network, and serves as a failover streaming backup download point together with the aforementioned backup server.

Each of the aforementioned backup servers and each P2P network can be defined as a peer-to-peer terminal, which has the functions of video and audio streaming, exchange, and storage, and is compatible with the aforementioned maintenance server of the data center network connection.

The aforementioned data center network can be set up as a multi-layer structure: most data centers serve as the edge data centers of the clients in the service area, which can provide local users with low communication delay video and audio streaming backup. The remaining data centers are configured at a higher level than the edge data centers in the architecture, which can absorb the occasional workload burst that the edge data centers cannot handle. All data centers support failover together, and its DNS mechanism directs the client seeking backup to the nearest available backup server or P2P download point.

Therefore, the characteristics of the present invention are as follows:

1. At the beginning of the video streaming, the aforementioned maintenance server confirms the location and number of clients in each region and configures the required backup servers in each data center, and then the main server will provide the streaming content of the client in a hierarchical structure Back up from top to bottom to the backup server or P2P peer of each data center.

2. When the client encounters a failure and the streaming is not continuous, the client connects to the data center network for failover service, and then the DNS mechanism of the data center network directs the client upwards to the nearest available device according to a hierarchical structure. Support server. Since it may take less than 2 seconds for the transfer client to connect to the backup server or P2P download point, the user can complete the connection to the backup server or P2P download point in other regions without knowing it and continue to watch the streaming content.

In the implementation of the present invention, it is possible that the local data center cannot provide enough backup servers or P2P download points for failover. In order to protect the rights and interests of users, in addition to the maintenance server will allocate the required backup servers in each data center according to the number of clients, each data center will also use the DNS mechanism to direct the client to a relatively low level when the backup server is insufficient. High-level data centers, such as the Google data center in Changhua, are higher-level data centers, while data centers in Taichung or Nantou are sub-level data centers. When the sub-level data centers cannot provide local client failover services, the client will be The DNS server of the local data center directed the Changhua data center for failover services. The foregoing implementation of the DNS-oriented mechanism can be combined with Software Defined Networking to re-plan the failover orientation in a centrally controlled manner using software programs.

)，那麼維護伺服器必須替每一個客戶端配置足以提供(x．log _f

)-bps影音串流的備援伺服器或P2P下載點。 The implementation method of maintaining server configuration redundancy is as follows: Assume that the streaming bandwidth required by each client is a fixed value of x-bps (bits per second), the time required for failover is T seconds, and the statistics show that the client The probability of failure in T seconds is f . To ensure that the failover success rate is at least (1-

), then the maintenance server must be configured for each client to provide (x.log _f

)-bps backup server or P2P download point for video streaming.

In the above implementation, at least two or more layers of data center networks are included, so that the edge data center can direct the unmanageable workload to the high-level data center, so as to reduce the load of the edge data center and strengthen the failure. The scalability of transfer services.

In the embodiment of the present invention, each backup server and each P2P download point are peers. Therefore, when the video streaming is in progress, the aforementioned maintenance server needs to ensure that each backup server is required The server and each P2P download point are in normal condition. If the backup server or P2P download point is not enough to support the existing customer base, the maintenance server will dynamically configure additional backup servers in the relevant data center.

(1): Video and audio streaming provider

(2): Data center network

(2’): High-level data center network

(21): Main server

(22): Maintain the server

(23): DNS server

(24): Redundant server

(3): P2P network

(4): Client

Figure 1 is a system diagram of the present invention; Figure 2 is a schematic diagram of the process of transmitting media audio and video of the present invention; Figure 3 is a state diagram of the present invention for backup maintenance; Figure 4 is the DNS server, A schematic diagram of the maintenance server assigning peers; Figure 5 is a schematic diagram of the failover process of the present invention.

Please refer to Figure 1. The architecture of the present invention can be built on all current data center networks. The diagram of the present invention takes CDN-P2P network as an example, but it is not limited to this and can also be used in CDN networks. , Academic network and similar network services.

Please refer to Figure 1. The CDN-P2P network architecture provided in this case is basically composed of the following equipment, including: an audio and video streaming provider (1), a data center network (2), multiple P2P Network (3) and multiple clients (4).

Please refer to Figure 1. The aforementioned video streaming provider (1) provides and uploads video streaming content to the aforementioned data center network (2). In other words, the video streaming provider (1) is like a live broadcast host or Live broadcast media companies provide live media video and audio, such as sports events, news, etc.

Please refer to Figure 1. The aforementioned data center network (2) includes a primary server (21), a maintenance server (22), at least one DNS server (23) and multiple backup servers (24) . The aforementioned main server (21) stores the video and audio streaming content from the video and audio streaming provider (1), and sends the streaming content to the backup server (24) where each client is located. The backup server (24) stores the backup of the video and audio streaming content from the main server (21). When the client (4) encounters network congestion, network delay, or the failure of the provider server, which causes the streaming to be interrupted and discontinuous, the backup server (24) will directly provide the streaming content backup to the client (4) Watch.

Please refer to Figure 1. The aforementioned P2P network (3) is composed of multiple clients (4) as a delivery network, and serves as a failover streaming backup download point together with the aforementioned backup server (24).

Please refer to Figure 1. Each of the aforementioned backup servers (24) and each P2P network (3) can be defined as a peer-to-peer terminal, which has the functions of video and audio streaming, exchange, and storage. And each backup server (24) and each P2P network (3) are connected with the maintenance server (22) of the aforementioned data center network (2).

Please refer to Figures 2 to 5, through the above system architecture description, the use of the present invention and obtainable effects are as follows: Please refer to Figure 2. At the beginning of the video streaming, the aforementioned maintenance server (22) confirms the location and number of clients (4) in each region and configures the required backup servers (24) in each data center. Then the main server (21) backs up the streaming content of the video and audio streaming provider (1) from top to bottom in a hierarchical structure to the backup server (24) or P2P network (3) of each data center. Wait. In other words, the client (4) is connected to the nearest backup server (24) with better viewing quality.

As shown in Figure 3, when the client (4) starts to watch media videos, the background of the client (4) will send back data to one of the peers or send the feedback data to the high-level data center network (2') , Any peer of the high-level P2P network (3'), but must comply with the operation of the parent stream and the child stream; the intention here is to reserve at least one storage point, if a data center network ( 2) When any peer of the architecture fails, the current progress can be continued, avoiding the problem of media video and audio interruption.

Please refer to Figures 4 and 5, the failover mechanism of the present invention is: when the client (4) encounters a failure and the stream is interrupted and discontinuous, the client (4) connects to the data center network (2) to seek failure Transfer services, and then the DNS server (23) of the data center network (2) directs the client (4) to the nearest available backup server (24) according to a hierarchical structure. Since it may take less than 2 seconds for the transfer client (4) to connect to the backup server or P2P download point, the client (4) can connect to the backup server (24) in other regions without knowing it. P2P network (3) Download point to continue watching the streaming content.

Please continue to refer to Figures 4 to 5. In the implementation of the present invention, it is possible that the local data center network (2) cannot provide a sufficient number of backup servers (24) or P2P network (3) download points. Perform a failover. In order to protect the rights and interests of users, each data center network (22) will configure the required backup server (24) in each data center network (2) according to the number of clients (4). 2) When the backup server (24) is insufficient, the client (24) will be directed to the higher-level data center network (2') through the DNS server (23). For example, the Google data center in Changhua is higher Hierarchical data center network (2'), and Taichung or The data center in the Nantou area is a sub-level data center network (2). When the sub-level data center network (2) cannot provide local client (4) failover services, the client (4) will be The DNS server (23) of the data center is directed to the higher-level data center network (2') of Changhua for failover services. The foregoing implementation of the DNS-oriented mechanism can be combined with Software Defined Networking to re-plan the failover orientation in a centrally controlled manner using software programs.

)，那麼維護伺服器(22)必須替每一個客戶端配置足以提供(x．log _f

)-bps影音串流的備援伺服器(24)或P2P網路(3)下載點。 It is worth mentioning that the implementation method of the aforementioned maintenance server (22) configuration backup is as follows: assume that the streaming bandwidth required by each client (4) is a fixed value of x-bps (bits per second), and the failure The time required for the transfer is T seconds, and the statistics show that the probability that the client (4) fails in T seconds is f . To ensure that the failover success rate is at least (1-

), then the maintenance server (22) must be configured to provide (x.log _f

)-bps video streaming backup server (24) or P2P network (3) download point.

In the embodiment of the present invention, at least two layers (ie, high-level and sub-level) or more than two levels of data center network (2) are included, so that the sub-level data center network (2) can handle tasks that cannot be handled The burst volume is directed to the high-level data center (2') to reduce the load of the sub-level data center (2) and strengthen the scalability of failover services.

In the embodiment of the present invention, each backup server and each P2P download point are peers. Therefore, when the video streaming is in progress, the aforementioned maintenance server needs to ensure that each backup server is required The server (24) and each P2P network (3) are in normal condition. If the backup server (24) or P2P network (3) is not enough to support the existing customer base, the maintenance server (22) will dynamically configure additional backup servers (24) in the relevant data center network (2) .

Therefore, the technical advantages provided by the present invention are roughly as follows: (1) It can be constructed using the current existing data center network (2), which can reduce the cost of setting up equipment for the industry; (2) It reduces the interruption of media video and audio due to failure Problem situation; (3) The industry can quickly perform system maintenance on the faulty peer.

In summary, the failover method of the hierarchical data center used in the video and audio streaming of the present invention has never been seen in books or publicly used, and it is in compliance with the requirements of an invention patent application.

If you need to clarify, the above are the technical principles immediately used in the implementation of the present invention. If the changes made according to the concept of the present invention, the functions produced by it still do not exceed the spirit covered by the specification and drawings, All should be within the scope of the present invention, and should be disclosed.

(1): Video and audio streaming provider

(2): Data center network

(2’): High-level data center network

(21): Main server

(22): Maintain the server

(23): DNS server

(24): Redundant server

(3): P2P network

(4): Client

Claims (5)

A failover method for audio and video streaming using hierarchical data centers. Any data center is constructed under the data center network architecture, including: an audio and video streaming provider (1), a data center network (2), and more A P2P network (3) and multiple clients (4); the aforementioned video and audio streaming provider (1) provides and uploads video and audio streaming content to the aforementioned data center network (2); the aforementioned data center network The road (2) includes a main server (21), a maintenance server (22), at least one DNS server (23) and multiple backup servers (24); the aforementioned main server (21) stores audio and video Streaming the audio and video streaming content from the streaming provider (1), and sending the streaming content to the backup server (24) where each client is located; the backup server (24) stores the aforementioned main server (21) Backup of incoming video and audio streaming content, and directly provide streaming content backup to the client (4) for viewing; the aforementioned P2P network (3) is composed of multiple clients (4) as a delivery network Used together with the aforementioned backup server (24) as a streaming backup download point for failover; each aforementioned backup server (24) and each P2P network (3) can be defined as a Peer-to-peer, which has the functions of video and audio streaming transmission, exchange, and storage, and maintenance server for each backup server (24) and each P2P network (3) and the aforementioned data center network (2) Server (22) connection; in the usual maintenance server (22), confirm the location and quantity of the client (4) in each region and configure the required backup server (24) in each data center, and then the main server ( 21) Backing up the streaming content of the audiovisual streaming provider (1) from top to bottom in a hierarchical structure to the backup server (24) of each data center or the peer of the P2P network (3); When the client (4) encounters a failure and the streaming is interrupted and discontinuous, the data center network (2) to which the client (4) is connected seeks failover services, and then the DNS of the data center network (2) The server (23) directs the client (4) to the nearest available backup server (24) according to a hierarchical structure. The failover method for audio and video streaming using hierarchical data center as described in claim 1, which is characterized in that when the lower-level data center network (2) of the hierarchical data center cannot provide a sufficient number of backup servers The server (24) or the P2P network (3) download point for failover, and the client (4) is directed to the higher-level data center network (2') through the DNS server (23). The failover method for audio and video streaming using hierarchical data center as described in claim 1, characterized in that: the implementation method of the aforementioned maintenance server (22) configuration backup is as follows: assuming that each client (4) needs The streaming bandwidth of is a fixed value of x-bps, the time required for failover is T seconds, and statistics show that the probability of failure of the aforementioned client (4) in T seconds is f ; if you want to ensure that the probability of successful failover is at least (1-

), the maintenance server (22) must be configured for the client (4) enough to provide (x.log _f

)-bps backup server (24) or P2P network (3) for audio and video streaming. The failover method for audio and video streaming using hierarchical data center as described in claim 1, characterized in that the hierarchical data center includes at least two or more data center networks (2' 2'), so that The sub-level data center network (2) can direct the unmanageable workload to the high-level data center (2') to reduce the load of the sub-level data center (2) and strengthen the scalability of failover services . The failover method for audio and video streaming using hierarchical data center as described in claim 1, characterized in that: if the aforementioned backup server (24) or P2P network (3) is not sufficient to support the existing client (4) ), the maintenance server (22) will dynamically configure an additional backup server (24) on the data center network (2).

Images