CN105872044A - Streaming media multi-level cache network acceleration system and method based on WebRTC - Google Patents

Abstract

The invention discloses a streaming media multi-level cache network acceleration system based on WebRTC. The system comprises a CDN network structure and an intra-group client P2P (Peer to Peer) network group structure. The intra-group client P2P network group structure obtains cache resources from edge cache nodes of the CDN network structure. The intra-group client P2P network group structure comprises a client in support of WebRTC, and a signaling server. The client in support of WebRTC applies a browser cache technique. The system is equipped with a center data source multi-point data cache structure, and the service capability reliability of the system is improved.

Description

Streaming media multi-level cache network acceleration system and method based on WebRTC

technical field

The invention relates to the field of network acceleration, in particular to a streaming media multi-level cache network acceleration system and method based on WebRTC.

Background technique

The P2P (Peer to Peer) technology is introduced into the CDN system, and the transmission mode between the content server and the user in each edge autonomous domain adopts the P2P mode. However, the controllability of the P2P network topology is poor, and the flow of data is disorderly. At the same time, each client node can join and exit the network at any time, which brings many challenges to the stability of the system. When some network users in the P2P group suddenly quit the P2P group, there may be a fault in the network streaming media resources in the P2P group, which affects the downloading process of the streaming media resources of the users. Among them, the P2P network topology is built based on the C/S architecture, but it is not so commonly used in the browser system B/S architecture where streaming media transmission is more frequent.

Contents of the invention

The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art, a streaming media multi-level cache network acceleration system and method based on WebRTC.

The purpose of the present invention is achieved through the following technical solutions:

A streaming media multi-level caching network acceleration system based on WebRTC, including a CDN network structure and an intra-group client P2P network group structure, the intra-group client P2P network group structure caches resources from the CDN network structure, the The client P2P network in the group includes a WebRTC-supporting client and a signaling server, and the WebRTC-supporting client applies browser caching technology.

The communication between the client supporting WebRTC and the signaling server and the mutual transmission of streaming media cache files between the clients are all implemented by using WebRTC communication.

The browser cache technology uses a cache mechanism defined by the HTTP protocol, and in HTTP 1.1, a Cache-Control strategy is used.

Another object of the present invention is achieved through the following technical solutions:

A WebRTC-based streaming media multi-level cache network acceleration method, specifically comprising the following steps:

S1. Building a CDN streaming media network structure according to the present invention by taking advantage of CDN network acceleration services, and distributing streaming media resources to resource servers and edge node cache servers.

S2. Since this system runs on a WebRTC-based browser, according to the real-time communication characteristics of WebRTC web pages, WebRTC technology supports point-to-point connection of clients; therefore, a signaling server is introduced when the client connects. It enables the clients in the group to be connected to each other, and realizes the client P2P network structure in the group.

S3. Since multiple clients with close geographic locations are likely to access the same edge node cache server, the edge node cache server and these clients can be divided into a group.

S4. Among the clients in the group that jointly access the same edge node cache server, these clients can not only obtain resources from the edge cache server, but also obtain resources from other clients in the group, so as to form a streaming media resource Third level cache.

Compared with the prior art, the present invention has the following advantages and beneficial effects:

1. The present invention greatly reduces the pressure on the central node data source and the edge cache node data source. If each service node of a traditional CDN wants to obtain the data source data of the central node, it needs to directly access the data source server, resulting in high pressure on the data source, high bandwidth consumption, and the lack of guarantee for the data link from the center to the edge. By organizing CDN service nodes in a P2P manner, nodes can share and cache data with each other through P2P, which greatly reduces the pressure on the central data source.

2. The central data source data of the present invention is backed up at multiple points to improve the reliability of system service capability. Different service nodes back up the central data at multiple points. This strategy improves the overall redundancy capability of the CDN system and the self-recovery capability of services, further enhancing system stability.

3. The present invention increases the number of nodes that can provide services and improves service flexibility. Share and cache data through P2P, so that the number of nodes that can provide services is greatly increased, and at the same time, the service of service nodes is more flexible and intelligent.

4. The present invention increases the scalability of the system and improves service efficiency. The underlying content distribution is realized by P2P technology, which greatly improves the scalability of the system. The overall system has good flexibility to cope with sudden and random user access and maintain good service efficiency.

5. The present invention improves the manageability of the network and avoids disordered flow. The system strictly limits the scope of P2P to the service area of a certain edge service node, avoiding the problems of backbone network congestion and traffic disorder caused by too many cross-regions and cross-ISPs caused by traditional P2P technology. Enhanced network manageability and service reliability.

Description of drawings

Fig. 1 is the structural diagram of the streaming media multi-level cache network acceleration system based on WebRTC of the present invention;

Fig. 2 is the structural representation of the JSEP adopted by the system described in Fig. 1;

Fig. 3 is a schematic diagram of the construction position of the STUN adopted by the system described in Fig. 1;

Fig. 4 is the P2P network structure schematic diagram that the signaling server that the system described in Fig. 1 adopts and the STUN server quickly set up;

FIG. 5 is a flow chart of the execution process of the browser cache adopted by the system shown in FIG. 1 .

detailed description

The present invention will be further described in detail below in conjunction with the embodiments and the accompanying drawings, but the embodiments of the present invention are not limited thereto.

The present invention provides a kind of method of streaming media multi-level cache technology, the method is:

1. With the advantages of CDN network acceleration service, build a CDN streaming media network structure belonging to the present invention, and distribute streaming media resources to resource servers and edge node cache servers.

2. Since this system runs on a WebRTC-based browser, according to the real-time communication characteristics of WebRTC web pages, WebRTC technology supports point-to-point connections of clients. So when the client connects, the signaling server is introduced. It enables the clients in the group to be connected to each other, and realizes the client P2P network structure in the group.

3. Since multiple clients with close geographic locations are likely to access the same edge node cache server, the edge node cache server and these clients can be divided into a group.

4. Among the clients in the group that jointly access the same edge node cache server, these clients can not only obtain resources from the edge cache server, but also obtain resources from other clients in the group, so as to form a stream media resources. Third level cache.

The present invention provides a streaming media multi-level caching technology system. As shown in FIG. 1 , the system is mainly composed of a CDN network structure and an intra-group client P2P network group structure.

The dashed part is the newly added server—the signaling server. Its function is that the client publishes and exchanges its own connection information by accessing the signaling server, and other clients can establish a connection with it through the signaling server to realize the interconnection between clients and establish a P2P network. After the introduction of the signaling server, the client can not only obtain resources directly from the site server, but also obtain resources from other connected clients. When the bandwidth of the site server is limited or the processing capacity is reduced, the website can still be accessed normally. Serve.

Through the comparison before and after the introduction of the acceleration service, it can be known that the network acceleration service consists of the following parts:

(1) Clients that support WebRTC

The network acceleration service here is built based on WebRTC, and the participating clients need to support WebRTC. Using the network communication part of WebRTC, each client can quickly form a P2P network. Whether it is a PC or a mobile terminal, Firefox and Chrome browsers provide most of the support for WebRTC, so most clients can use WebRTC without any installation.

(2) Signaling server

Using WebRTC to build the client's P2P network, the clients need to exchange their own information to connect together. At this time, signaling service (Signaling) is required. Signaling is a process of coordinating communication. In order to form a session between clients, clients need to exchange the following signaling information: initiate and close a call control information; error information; metadata information; keys to ensure communication security; network Data (such as the IP address and port of the host on the external network). The signaling handling on the client side requires a way to pass information back and forth, this mechanism is not defined by WebRTC, you need to create it yourself. The structure of JSEP is shown in Figure 2:

The structure of JSEP prevents the browser from saving state information. If the browser is allowed to save the signaling state, the signaling will disappear every time the page is reloaded. A better solution is to save the signaling state with a server, i.e. use a signaling server. WebRTC's signaling server does not spend too much memory and processes, and only needs to forward information and maintain very little session state data, so it is even possible to open a process on the current server to be responsible for signaling processing. But the signaling mechanism can be used not only to exchange session metadata, but also to transmit application data. It is essentially an information service, so it is regarded as a separate server here.

(3) STUN/TURN server

After the signaling interaction is completed, the client will try to connect directly to other clients. In the simple web world, each WebRTC client has a unique address, so that they can communicate directly, but in reality, most devices are behind one or more NAT layers, using proxies or firewalls. In order to overcome the complex network environment in the real world, it is necessary to use the STUN server to obtain the external network address, or use the TURN server, which adds a relay function to the STUN server. The location of STUN is shown in Figure 3:

The STUN service is set up on the external network, obtains the IP and port of the client running behind the NAT, and then returns the address. The role of STUN is to discover the external network IP and port of the client, and the subsequent execution steps are the execution process of the signaling service. Therefore, the STUN server does not need to store too much content or do too much work, and a simple STUN server can handle a large number of requests. When the simple STUN is invalid, it will turn to TURN. The role of the TURN server is to relay data. It is responsible for transferring audio and video, data streams, etc. between the two ends, but not sending data. TURN has a public address, so every client can access it (regardless of using a proxy or firewall). The task of TURN is simple, but data transfer will consume a lot of bandwidth, so TURN needs to be strong enough.

For normal use, you can use the STUN server provided by Google, but if it is a practical product, you need to build a STUN server yourself to reduce the risk of using third-party services.

The WebRTC client, signaling server and STUN server can quickly establish a P2P network to achieve the effect of network acceleration. The complete picture is shown in Figure 4:

2) Key technical points of the present invention

The three key points of this three-level cache structure are signaling server, client and browser cache technology.

(1) Clients that support WebRTC

The network acceleration service here is built based on WebRTC, and the participating clients need to support WebRTC. Using WebRTC webpage real-time communication technology can enable clients to transmit streaming media resources to each other. Whether it is a PC or a mobile terminal, Firefox and Chrome browsers provide most of the support for WebRTC, so most clients can use WebRTC without any installation.

(2) Signaling server

Using WebRTC to build the client, the clients need to exchange their own information to form a P2P network group in order to establish a connection with each other. At this time, signaling service (Signaling) is required. Signaling is a process of coordinating communication. In order to form a session between clients, clients need to exchange the following signaling information: initiate and close a call control information; error information; metadata information; keys to ensure communication security; network Data (such as the IP address and port of the host on the external network). The signaling handling on the client side requires a way to pass information back and forth, this mechanism is not defined by WebRTC, you need to create it yourself.

The communication between these clients and the signaling server and the mutual transmission of streaming media cache files between clients are all realized by using WebRTC communication. WebRTC provides three APIs that can be called at the application level, namely MediaStream, which is used to access the camera and microphone of the device to obtain synchronous video and audio streams; RTCPeerConnection, which is used to build stable and efficient stream transmission between point-to-point; RTCDataChannel, Establish a low-latency, high-throughput channel between point-to-point, which can be used to transmit any type of data.

(3) Browser caching technology

After the client-side communication technology is perfected, client-side caching, that is, browser caching technology, becomes very important. All browsers provide cache, which caches pictures, text, and script files of visited sites, and even caches video files requested by page players. The browser's cache mechanism uses the cache mechanism defined by the HTTP protocol. In HTTP 1.1, the Cache-Control strategy is used. The principle is to control whether the browser re-sends a request to the server to obtain data or reads data from the browser cache by specifying the validity period of the cached resource. Cache-Control has multiple setting options, and the optional values of Cache-Control in the HTTP protocol header are shown in the table:

Table 1

The fields Last-Modified/If-Modified-Since and Etag/If-None-Match also need to be used in conjunction with Cache-Control. The description of these four fields is shown in the table:

Table 2

Etag is the unique identifier of the resource on the server side, which is automatically generated by the server and can control the cache content more precisely. If you use Last-Modified and Etag at the same time, the server will verify Etag first, and compare Last-Modified when Etag is consistent.

The flow chart of the execution process of browser caching is shown in Figure 5. After these browsers cache streaming media files, the protocol used for file transfer between browsers is HLS (HTTP Live Streaming). HLS is an HTTP-based streaming media transmission protocol proposed by Apple. Its working principle is to split the entire stream into small files. This file is downloaded through HTTP. Instead of downloading all fragments each time, it downloads a part. When streaming, the client can choose to download the same resource at different rates from multiple different sources, and the streaming session can accommodate different data rates. Before streaming starts, the client will first download an extended M3U playlist file (m3u8 index file), which contains metadata and is used to find media streams available in the network. Since HLS only requests HTTP packets, it is easy to use CDN to transmit and distribute streaming media, and unlike real-time transport protocols, HLS can pass through any firewall or proxy server that allows HTTP data to pass through.

Under this CDN network structure, by adding edge cache servers and internal DNS responsible for scheduling, it is reasonable for users to request the cache server that is always the closest, which effectively reduces the access pressure of the central station. Even if a node encounters access delay or is attacked, it can quickly respond to user requests through redundant nodes. Due to the CDN architecture, it has the characteristics of local acceleration, remote acceleration, bandwidth optimization, server mirroring and attack resistance. The anti-attack feature is that CDN, a large cluster, is widely distributed and has an intelligent redundancy mechanism, which can effectively prevent hacking and various DDoS attacks from affecting service access, ensure data security, and provide better service quality .

Intra-group client P2P network group structure sharing the same edge node server. The newly added server in this network structure is the signaling server. Its function is that the client publishes and exchanges its own connection information by accessing the signaling server, and other clients can establish a connection with it through the signaling server to realize the interconnection between clients. After the signaling server is introduced, the clients in the group can be connected to each other, and a P2P network is realized in the group. Clients in the group can obtain resources not only from the edge cache server, but also from other clients in the group. By using the client P2P network acceleration technology in the group to integrate with the CDN network structure, the client streaming media resources are formed. The three-level caching strategy allows the client to access site resources from the nearest and multi-layer caching nodes to realize streaming media network acceleration services. In the P2P network of the clients in the group, since the clients can transfer resources to each other, when multiple clients upload resources, the client download can almost reach the maximum speed of its own bandwidth without being limited by the bandwidth of the server. Moreover, in the P2P network structure within this group, the client can obtain the connection information of other clients by accessing the signaling server, so as to reasonably select the nearest client for connection. In this way, the access pressure to the edge node server can be reduced.

WebRTC to build a client-side P2P network implementation:

(1) Clients that support WebRTC

The network acceleration service here is built based on WebRTC, and the participating clients need to support WebRTC. Using WebRTC webpage real-time communication technology can enable clients to transmit streaming media resources to each other. Whether it is a PC or a mobile terminal, Firefox and Chrome browsers provide most of the support for WebRTC, so most clients can use WebRTC without any installation.

(2) Signaling server

Using WebRTC to build the client, the clients need to exchange their own information to form a P2P network group in order to establish a connection with each other. At this time, signaling service (Signaling) is required. Signaling is a process of coordinating communication. In order to form a session between clients, clients need to exchange the following signaling information: initiate and close a call control information; error information; metadata information; keys to ensure communication security; network Data (such as the IP address and port of the host on the external network). The signaling handling on the client side requires a way to pass information back and forth, this mechanism is not defined by WebRTC, you need to create it yourself. The structure of JSEP is shown in Figure 2: the structure of JSEP prevents the browser from saving state information. If the browser is allowed to save the signaling state, the signaling will disappear every time the page is reloaded. A better solution is to save the signaling state with a server, i.e. use a signaling server. WebRTC's signaling server does not spend too much memory and processes, and only needs to forward information and maintain very little session state data, so it is even possible to open a process on the current server to be responsible for signaling processing. But the signaling mechanism can be used not only to exchange session metadata, but also to transmit application data. It is essentially an information service, so it is regarded as a separate server here.

The above-mentioned embodiment is a preferred embodiment of the present invention, but the embodiment of the present invention is not limited by the above-mentioned embodiment, and any other changes, modifications, substitutions, combinations, Simplifications should be equivalent replacement methods, and all are included in the protection scope of the present invention.

Claims (4)

1. the Streaming Media multi-level buffer network acceleration system of sing on web RTC, it is characterised in that include CDN structure and group Interior client P2P group of networks structure, in described group, client P2P group of networks structure caches money from described CDN structure Source, in described group, client P2P network includes client and signal server, the described support WebRTC supporting WebRTC Client application browser rs cache technology.

The Streaming Media multi-level buffer network acceleration system of sing on web RTC the most according to claim 1, it is characterised in that Communicate between client and the described signal server of described support WebRTC and to realize flow medium buffer file between client mutual Pass is all to use WebRTC communication to realize.

The Streaming Media multi-level buffer network acceleration system of sing on web RTC the most according to claim 1, it is characterised in that The caching mechanism that described browser rs cache technology uses http protocol to define, in HTTP 1.1, employs Cache-Control strategy.

4. the network acceleration side of Streaming Media multi-level buffer network acceleration system based on the sing on web RTC described in claim 1-3 Method, it is characterised in that specifically include following steps:

S1, the advantage serviced by CDN acceleration, build a CDN stream media network structure belonging to the present invention, and Streaming media resource is distributed in Resource Server and fringe node caching server.

S2, due to native system be on the browser of sing on web RTC run, according to the spy of WebRTC webpage real-time Communication for Power Property, the point-to-point connection of WebRTC technical support client；So client connects when, introducing signal server. Make the client in group can realize connecting each other, in group, achieve client P2P network structure.

S3, same fringe node caching server may be accessed owing to multiple clients that geographical position is close are the biggest, it is possible to Fringe node caching server and these clients are divided into a group.

In S4, client in the common group accessing same fringe node caching server, these clients can not only be from edge cache service Device obtains resource, it is also possible to other clients in group obtain resource, forms the third level caching of streaming media resource with this.