CN100558042C - A P2P live broadcast method based on super nodes - Google Patents

Abstract

The present invention relates to a P2P live broadcasting method based on a super node, the method comprising: a node sends a node joining request information of a desired program to a server, and the server processes the node joining request message to complete the joining of the node into a live streaming network; and the After a node successfully joins the live streaming network, the node chooses to establish a partnership with other nodes that are physically close to itself, and then transmits data. Therefore, the present invention enables data to be transmitted between similar common nodes on the underlying physical network, reducing the number of hops and time delays in the data forwarding process; burden, reducing bandwidth consumption; at the same time, the present invention separates node management from data transmission, making maintenance easier.

Description

A P2P live broadcast method based on super nodes

technical field

The invention relates to the field of Internet communication, in particular to a peer-to-peer (P2P) live broadcast method based on super nodes to realize real-time media transmission.

Background technique

Streaming media technology is a technology that uses streaming to play audio, video and other multimedia data on the Internet. Users can watch while downloading, without downloading the entire file before watching. What stream media technology adopts at first is traditional client/server (C/S) mode, and multimedia data is stored on the server, and any client wants to obtain media data, must obtain a complete copy from server. In this mode, the server must send exactly the same data to each user. As the number of users increases, sending huge multimedia data to a large number of users will overwhelm the server and cause server-side congestion.

In order to alleviate the pressure on the server in the traditional C/S mode, proxy server technology is applied to network media publishing. The proxy server stores some frequently accessed data in the local storage. When the user accesses through the proxy server, if the cache of the proxy server has the data the user needs, there is no need to access the remote server, and only the local cache serves the user. . The proxy server shares part of user access, effectively reducing the pressure on the main server, improving user response speed, and saving network resources. However, the deployment of proxy servers will increase the hardware cost of media publishing, and it is also a difficult problem to keep the content in the proxy cache and the content of the main server synchronized.

Content Distribution Network (CDN) is a virtual network built on the Internet and composed of node server groups distributed in different regions. The core idea of CDN is to push the content from the center to the edge close to the user, which not only effectively improves the performance of the user's access to content, but also effectively reduces the pressure on the central equipment and the backbone network. Through CDN, the content service can be changed from the original single center structure to a distributed structure. The CDN system effectively solves the content synchronization problem between the proxy cache and the main server in the proxy server technology. However, CDN is still a distributed media service technology platform based on C/S structure, which is closely connected with regional control, so it is difficult to reduce the cost of its expansion. The adaptability and fault tolerance of CDN technology to burst traffic during peak periods are still There are certain defects, such as the bandwidth of the streaming media server is large, and the processing capacity of the server is high.

The above technologies are all developed on the basic structure of C/S, so the problems of system bottleneck and single point of failure caused by the server itself cannot be avoided. In order to solve the problems existing in the traditional C/S structure, P2P technology has been widely developed. In the P2P system, it is no longer like the traditional C/S structure that only the server provides data for users to download. Users can also download data while downloading. Provide services to other users, and in real applications more data is transmitted between users. With the popularity of P2P downloading technology represented by BT, P2P technology has attracted more and more attention.

However, the existing P2P streaming media technology does not consider the topology matching between the overlay network and the underlying physical network. The nodes are very similar in the logical overlay network, but they may be far apart in the physical network. and the speed of transmission will be greatly affected, and the quality of service obtained by users cannot be effectively guaranteed. At the same time, the server in the system must also control and manage the entire system in addition to distributing media data. Some systems specifically provide control and management servers. , but this increases the input cost of hardware.

Contents of the invention

The purpose of the present invention is to provide a P2P live broadcast method based on super nodes in order to solve the topology matching problem of the overlay network and the underlying physical network in the prior art.

In order to achieve the above object, the present invention provides a P2P live broadcast method based on super nodes, the method comprising:

The node sends the node joining request information of the desired program to the server, and the server goes to the program list to match the super node according to the program that the node wants to watch in the node joining request information;

The process of matching the super node is: compare the flag value of the node with the flag value of the super node in the server list, if they are the same, the node and the super node are in the same physical network or are physically close to each other match;

The server returns the information of the successfully matched super node to the node, and the node sends the node joining request message to the super node according to the information of the super node, and the super node responds to the node joining request Message to join the live streaming network with the node as a normal node, and return the information of the managed normal node to the node; and

After the node successfully joins the live streaming network, the node chooses to establish a partnership with other nodes that are physically close to itself, and then transmits data.

Further, the super node-based P2P live broadcast method further includes: the node withdraws from the streaming media live broadcast network.

Further, the process for the node to choose to establish a partnership with other nodes that are physically close to itself is specifically as follows:

The node sends a partnership establishment request to the super node;

The super node returns its own partner information according to the partnership establishment request;

After the node receives the partner information returned by the super node, it selects a node that is physically close to itself and requests to be its partner;

If the requested node judges that there is enough bandwidth and the load does not exceed the limit value, it will establish a partnership with the requesting node.

Further, between the node sending a partnership establishment request to the super node and the super node returning its own partner information according to the partnership establishment request:

When the supernode receives the request to establish a partnership relationship, it determines that the load does not exceed the limit value after establishing a partnership relationship with the node, and the available bandwidth meets the requirements of the node, and then establishes a partnership relationship with the node.

Further, after the requested node judges that there is sufficient bandwidth and the load does not exceed a limit value, and establishes a partnership with the requesting node, the following steps are also included:

The node judges whether the number of its own partners is sufficient to meet its own needs. If the number of partners is still relatively small, the node sends a request to establish a partnership relationship with a node with a small delay between itself and itself, and establishes a partnership list.

Further, the process of the node withdrawing from the live streaming network includes the following steps:

The node sends a leaving message to its partner and the super node to which it belongs, notifying the partner and the super node that it will leave the live streaming network;

The partner deletes the information of the node from the partner list after receiving the leaving message, and the super node deletes the information of the node from the managed node list after receiving the leaving message, and simultaneously notifies Other normal nodes under management in order to update their node lists.

Further, the server processes the node joining request message to complete the node joining the streaming media live network specifically as follows:

The server adds the node as a super node to the list of super nodes, and notifies the node to join the live streaming network as a super node; and

The node joins the live streaming network as a super node.

Further, the process for the node to choose to establish a partnership with other nodes that are physically close to itself is specifically as follows:

When the node acts as a super node, the super node will add the node with larger available bandwidth and better performance among the ordinary nodes it manages to its partner list, and establish a partnership with it.

Further, the process for the node to choose to establish a partnership with other nodes that are physically close to itself is specifically as follows:

When the node acts as a super node, the super node requests the nodes in the network to establish a partnership;

The requested node judges that its load does not exceed the limit value, and the bandwidth is relatively sufficient, and establishes a partnership with the super node;

The requested node returns the information that it has established a partnership with itself to the super node;

From the returned information, the super node selects a node with a small delay between itself and itself to send a request for establishing a partnership relationship, and establishes a partnership list.

Further, the process of the super node withdrawing from the live streaming network includes the following steps:

The super node sends a leave message to its partner and server to notify the partner and server that it will leave the network, the partner deletes the information of the super node from the partner list after receiving the leave message, and the server receives the After leaving the message, select one with good performance as the new super node from the list of ordinary nodes under the management of the super node;

The server sends the information of other ordinary nodes to the new super node so that it can build the node list;

The super node sends a notification message to the normal nodes informing them that they are managed by a new super node;

Ordinary nodes update the super node information in their own node list according to the notification message.

The present invention adopts the two-layer management mode of the server and the super node, fully utilizes the performance of the client, and tries to transmit the data between common nodes that are close to each other on the underlying physical network, reducing the burden on the server and the super node, and reducing the bandwidth consumption, and at the same time, the present invention separates node management from data transmission, making maintenance easier. In the present invention, both super nodes and ordinary nodes only need to manage the node information they need, and can dynamically adapt to changes in the network, and have good scalability.

Description of drawings

Fig. 1 is the structural diagram of the P2P live broadcast system that the present invention is based on the P2P live broadcast method of super node;

Fig. 2 is the flowchart of the P2P live broadcasting method based on the super node of the present invention;

Fig. 3 is the flow chart that introduces two kinds of situations of node joining stage in detail;

Figure 4 is a description of the process of establishing a partnership between a super node and an ordinary node;

Fig. 5 is a flow chart of data transmission between a node and a partner node.

Detailed ways

As shown in Figure 1, it is the P2P live broadcast system used by the P2P live broadcast method based on super nodes in the present invention, the P2P live broadcast system of the present invention is composed of server 1, super node 2 (Super Node, SN) and ordinary node 3 (Ordinary Node, ON) consists of a three-tier architecture. Wherein, as shown in the figure, the common node 3 may be located in a local area network.

The server 1 mainly completes the distribution of media data, manages the super node 2, and processes the request information function of the new node. Data distribution performs segmentation processing on the video streaming data, records relevant segmentation information, and transmits the streaming data to the requesting node. Managing supernodes includes communicating with supernodes and managing the list of supernodes. The server maintains a list of supernodes. This list first classifies nodes according to the programs they watch. The nodes are classified according to their physical location, and those nodes with good performance are set as super nodes, and each super node manages those ordinary nodes that are close to it in physical location. The server dynamically manages this list, regularly selects super nodes from all nodes to replace them according to the performance of the nodes, dynamically updates the list of super nodes and maintains the information of all nodes according to the addition and exit of nodes. The server will control the load according to its own load so that its load will not be too heavy, so as to effectively provide services for the system. For the request to join a new node, the server will match the programs watched by the node and the geographical location, return a matching super node to the new node, and notify it to request the returned super node to join the overlay network managed by the super node among them. If there is no matching super node, the server will notify it to join the network as a super node.

The super node 2 mainly completes the data transmission, manages the ordinary node 3, manages the partner list, and processes the request information for new nodes. Data transmission is similar to the data distribution function of the server, the difference is that supernodes, in addition to providing data transmission to other nodes, also download and play data like ordinary clients. A super node maintains two tables, one is a list of ordinary nodes managed by itself, and the other is a list of partners. Partners refer to those nodes that watch the same program as themselves and exchange data with themselves. For the ordinary node list, the super node will dynamically update according to the node's joining and exiting operations, and notify the server of the update information so that the server can update the node list. The super node will also update the partner list when necessary. For example, if it does not find a certain data segment within a certain period of time, it will think that its partner node does not have the data it requested. It is necessary to update the partner list. When a new node sends a request to the super node, the super node will add the relevant information of the node to its own list, and notify the requesting node to join the overlay network managed by itself as a normal node.

Common node 3 mainly completes data transmission, partner node list management functions, and communication management functions. Among them, the management functions of data transmission and partner node list are similar to those of super nodes. Communication management is mainly to communicate with the super node and notify the super node of some update information. The joining and leaving of nodes in the P2P network is quite frequent. In order to ensure that the nodes do exist in the network, each ordinary node will regularly send its own survival information to the super node. The node sends the information of leaving the network, so that the super node can update the node list in time to manage the node effectively.

FIG. 2 is a flow chart of the super node-based P2P live broadcast method of the present invention. As shown in the figure, the method includes step 10: the node sends the node joining request information of the desired program to the server, and the server processes the node joining request message to complete the joining of the node into the streaming media live broadcast network; and step 20: the successful joining of the node After arriving at the live streaming network, the node chooses to establish a partnership with other nodes that are physically close to itself, and then transmits data.

That is, the P2P live broadcast method based on super nodes can be divided into three stages at first. The first stage is the node joining stage, which is the content shown in step 10. The node joining stage includes two situations, one is that the server joins the request information according to the node The program to be watched by the node in the program list matches the super node; the server returns the information of the super node that is successfully matched to the node, and the node sends the super node information to the super node according to the information of the super node. The node sends the node joining request message, and the super node responds to the node joining request message to join the node as a normal node into the streaming media live broadcast network, and returns the information of the managed common node to the node. The other is: the server adds the node as a super node to the super node list, and notifies the node to join the live streaming network as a super node; and the live streaming network that the node joins as a super node in the network. The second phase is the partnership establishment phase, and the third phase is the data transfer phase. The second stage and the third stage are the contents shown in step 20 .

Referring to FIG. 3 for details, two situations in the node joining phase are introduced in detail below. as the picture shows,

Step 301: the node sends node joining request information to the server;

The content of the request information includes the IP address of the requesting node, the listening port number, the hashID value of the node, the bandwidth of the node, the speed of the processor, the memory, and the program requested by the node. Among them, the hash ID value of the node is calculated by the IP address and port through the hash algorithm, which is used to uniquely identify the node, and the bandwidth, processor speed and memory of the node are used to calculate the performance of the node.

Step 302: matching with super nodes;

The server matches the super node with the corresponding program list according to the program the node wants to watch. The global Internet is divided into multiple levels according to the network service provider information and geographic information, respectively representing ISP, city, community (or university), and building. These information can be obtained from the free or commercial IP address to network location information conversion database. According to the obtained information, a network location flag value is generated for each node. The length of this flag value is 40 bits, of which 0 to 15 The 16th to 31st digits represent the city code, and the 32nd to 39th digits represent the ISP number. The matching process is to compare the flag value of the request node with the flag value of the super node in the server list. If they are the same, it is considered that the request node and the super node are in the same physical network or they are physically close. If the matching is successful, go to step 303, otherwise execute step 307;

Step 303: return the matching super node information;

The server will return the successfully matched super node information to the requesting node. The returned information includes the super node's hash ID value, IP address, listening port, etc. At the same time, the server adds the information of the request node to the super node list in the server.

Step 304: the node sends a node joining request message to the super node;

The requesting node obtains the IP address and port number of the super node from the super node information returned by the server, and sends a request message to the super node, and the content of the message is the same as that sent to the server.

Step 305: the super node processes the request:

The super node judges the network environment where the node is located. If it detects that the requesting node is in the public network, that is, step 3051: whether it is behind a firewall, if it is in the public network, then perform step 3053: the super node adds the node information that initiated the request to its own In the node list, otherwise the node is a node behind the firewall, execute step 3052: the supernode notifies it to perform NAT traversal, opens the relevant service port, and judges whether the NAT traversal is successful, if successful, then executes step 3053, and the supernode will initiate The information of the requesting node is added to its own node list, otherwise step 3054 is executed, the requesting node is refused to join, and the joining fails.

Step 306: return common node information;

The super node returns the information of the ordinary nodes it manages to the requesting node, and the information of these ordinary nodes is in preparation for establishing a partnership in the future, and the requesting node joins the network as an ordinary node, and executes step 309;

Step 307: the server adds the requesting node as a super node to the list of super nodes, and notifies the requesting node to join the network as a super node;

Step 308: requesting the node to join the network as a super node;

Step 309: the joining process ends.

At the beginning, there is only a server in the network. When the first node requests to join the network, the server will add the node as a super node to the node list according to the program requested by the node, and notify the node to join the network as a super node , at this time, the media data can only be transmitted between the server and this node. The process of joining new nodes in the future is the same.

The second phase of this P2P live broadcast method - the partnership establishment phase - is described below:

When a node successfully joins the network, the node needs to establish a partnership before transmitting data. The so-called partner refers to the node that watches the same program as itself and is willing to exchange data with itself. In the present invention, the node does not establish a partnership with all nodes watching the same program in the network, because the scale of the network is very large , There are many nodes watching the same program in the network, and a node does not need to exchange data with all nodes to meet its own needs. In the process of establishing a partnership, nodes will use the selection mechanism to try to select those nodes that are physically close to themselves as their partners for data exchange, which greatly reduces the pressure on servers and super nodes, and reduces the underlying physical transmission. bandwidth and latency.

In the above process, the node that initiates the node joining request message can join in the mode of super node or ordinary node when joining the live streaming network. When the requesting node joins the live streaming network, it needs to establish a partner for data exchange. The process of establishing a partnership between a super node and an ordinary node is described in detail below, as shown in Figure 4:

Step 400: Judging the type of nodes into ordinary nodes and super nodes, the establishment process of ordinary nodes includes:

Step 401: Send a request to establish a partnership.

The node first sends a request for establishing a partnership relationship to the super node to which it belongs. Because the server has classified the nodes in the network according to the programs, ordinary nodes and super nodes under the management of the same super node watch the same program, so ordinary nodes can directly request the super node to establish a partnership .

Step 402: The super node judges whether to establish a partnership

Supernodes are ordinary users, and they have to do some other things besides providing services, so all the hardware resources of supernodes cannot be used to provide services. The goal of this invention is to enable supernodes to provide services without affecting The normal work of the super node makes the super node feel that it does not consume too many resources. Therefore, in practice, the load of super nodes should be effectively controlled, so that the load value of super nodes can be limited within a certain range.

When the super node receives the request to establish a partnership, it judges its own load and available bandwidth. If the load does not exceed the limit value after establishing a partnership with the requesting node, and the available bandwidth is sufficient to meet the needs of the node, then execute step 403 : establish a partnership relationship with the requesting node, otherwise do not establish a partnership relationship with the node, and execute step 404 .

Step 404: The super node returns its partner information

Regardless of whether the requesting node has established a partnership with its super node or not, the number of partners of the requesting node is not enough to meet the demand at this time, so the super node will return the information that it has established a partnership with itself to the requesting node. Nodes that establish partnerships may include servers, other super nodes, and ordinary nodes managed by themselves. In order to ensure that data is exchanged between nodes that are physically similar to each other, super nodes will preferentially select ordinary nodes managed by themselves to return to requesting nodes.

Step 405: After receiving the partner information returned by the super node, the node selects the requested node that is physically close to itself as a partner;

After receiving the node information returned by the super node, the requesting node does not immediately send them a request to establish a partnership, but makes a certain selection to ensure that the partner is similar to itself. The selection process is to request the node to check the delay between itself and these nodes, exclude those nodes with relatively large delay, and send a request to establish a partnership to the remaining nodes.

Step 406: The requested node determines whether to establish a partnership

Like super nodes, it is also necessary to control the load of ordinary nodes. The node receiving the request also checks its own load and bandwidth. If there is enough bandwidth and the load does not exceed the limit value, it will establish a partnership with the requesting node. Otherwise, return the information of the node that has established partnership with itself to the requesting node.

In the above process, ordinary nodes can complete the partnership establishment.

Step 407: Whether the number of partners is sufficient

Through the above process, the requesting node has established partnerships with some nodes, and these partners are physically close to the requesting node. At this time, the requesting node will judge whether the number of its partners is sufficient to meet its own needs. If the number of partners is still relatively small, the requesting node will send the establishment of a partner to those other partners of the super node to which it belongs. relationship request.

Step 408: Build partnership list

Repeating the above process, that is, repeatedly executing steps 401 to 407, the requesting node finally forms its own partnership list, and these partners are physically close, so that the transmission delay can be relatively small.

Among them, the super node establishment process includes:

There are many similarities between the process of establishing a partnership between a super node and that of an ordinary node. These places will not be described in detail, but the differences will be described. The specific establishment process includes two types, one with The node managed by oneself establishes a partnership; the other is: establishes a partnership with a certain node in the network, as shown in Figure 4, including the following steps:

Step 410: Establish partnership with self-managed nodes

The super node will add those nodes with relatively large available bandwidth and good performance among the ordinary nodes it manages to its partner list, and establish a partnership with it.

Step 411: Send a request to establish a partnership

The super node first requests a certain node in the network to establish a partnership. For the super node, this node is a server at the beginning. For the method of the present invention, the server mainly transmits data to the super node, and the data to the common node Transmissions are minimal.

Step 412: Determine whether to establish a partnership

The requested node judges its own load and bandwidth. If the load does not exceed the limit value and the bandwidth is relatively sufficient, then execute step 415. The requested node establishes a partnership with the super node; if the load is relatively large or the bandwidth is not enough, it does not establish a partnership with the super node. partnership, go to step 413.

Step 413: The requested node returns its partner information

The requested node returns the information that it has established a partnership with itself to the super node.

Step 414: Among the node information returned by the super node, exclude those with a relatively large delay with itself, and send a request to establish a partnership to the remaining nodes. The specific process is the same as steps 405-408, forming a partnership list, and repeating the above process , supernodes eventually form their own list of partnerships that are physically close.

After the partnership is established, the node can request data from the partner node for data exchange. Referring to FIG. 5 , it shows a flow chart of the third stage—the data transmission stage.

The present invention adopts buffer map BM (Buffer Map) mode for the management of data, and this is a kind of data management mode generally adopted by existing P2P streaming media. In this method, the video streaming data is divided into data fragments of the same size, and a cache map is used to store the information of the relevant fragments, indicating the ownership of a certain data fragment. Nodes exchange data by checking each other's BM. ongoing. In the specific implementation, each node will cache data for a period of time. BM uses a bit to indicate the status of the node's possession of a data segment. 1 indicates that the node owns the data segment, and 0 indicates that the data segment does not exist.

At this stage, the node will query its partner node for data fragments according to the required streaming media data fragments. The requested partner node will check whether it owns the data fragment according to the information in BM. If it has, it will return information to notify the requesting node. I own this data fragment. In fact, there are often more than one partner nodes who own a certain data fragment. In this way, the node will choose which node to obtain this data fragment from. Nodes with low latency to exchange data.

As shown in Figure 5, the data transmission flow chart between the node and the partner node, the specific steps at this stage are as follows:

Step 501: Form a collection of data fragments

After the partnership is established, the node starts to request the program. At this time, the node will form a collection of information related to the required media data fragments, because the data fragments required by the node will continue to change as time progresses, and the obtained data fragment information will be deleted from this collection, and new request data fragment information will be added to this collection.

Step 502: The node initiates a data request

The node first selects a piece of data in this collection, and then initiates a data request to search among the partner nodes.

Step 503: The partner node returns a response

The partner node checks its own BM according to the data fragment requested by the node. If it finds that it owns the data fragment, it sends a response to the requesting node, telling the node that it owns the data fragment. If it does not have the data fragment, it returns that it does not own the fragment. the response to.

Step 504: The node selects a partner node

The requesting node judges the number of nodes that own the requested data segment. If the number of nodes that own the segment is greater than one, execute step 505 to select the node with the smallest delay with itself among these nodes. If there is only one node, go to step 506; The node of this data fragment goes to step 508 .

Step 506: Nodes perform data transmission

The requesting node sends a request for data transmission to the selected node, and the selected node transmits the data fragment it owns to the data requesting node.

Step 507: The node performs an update set

After receiving the data segment, the requesting node deletes the information of the data segment from the collection, and adds the information of the new required data segment to the collection according to the time information, and then goes to step 502 .

Step 508: Processing of not finding the data fragment

For each data segment, the present invention records the time when the query request is first issued, so as to track the elapsed time of the data segment from the initial query. When the owner of the data fragment is not found, judge whether this time exceeds the specified time T such as 20 seconds, if it is found that this value is exceeded, then execute step 508 and consider that there is no owner of this data fragment among the partners of the node, so the node proceeds The update of the partner list, then go to step 502, if this value is not exceeded, then directly go to step 502 and wait for the next query.

The present invention shares the control load of the server by introducing a super node, so that the server can better provide effective media data distribution for the system, and the super node in the system will manage those nodes close to itself on the physical network, so that the When the number of users is relatively large, users basically obtain data from those users who are physically close to themselves, thereby reducing the connection pressure on the server and the delay of data transmission, which can greatly reduce the consumption of underlying network bandwidth and improve service quality performance.

In addition to the above three stages, the super node-based P2P live broadcast method of the present invention also includes the stage that the node leaves the streaming media live broadcast network. The following will introduce the detailed process of the node leaving the live streaming network stage. The process includes: ordinary nodes and super nodes leave the network normally or abnormally.

The normal leaving process of an ordinary node is simple. The leaving node sends a leaving message to its partner and its super node, informing them that it will leave the live streaming network. After receiving this message, the partner node transfers the information of the node from its own After receiving this message, the supernode deletes the information of the node from the managed node list, and at the same time notifies other ordinary nodes under management so that they can update their own node list. At this time, the super node does not send the node's leaving information to the server, because the node can leave at any time in the P2P network, and the node leaves more frequently. In order to reduce the server's node management overhead The update information of the node list is sent at the same time as the information.

When a supernode leaves normally, first its partners and servers send a leave message to inform them that they are about to leave the network. After receiving this message, the partner deletes the information of the supernode from the partner list. After receiving this message, the server will Select a good performance as a new super node from the ordinary node list under the super node management to take over these ordinary nodes, and notify the selected node to convert to a super node. The server sends the information of other ordinary nodes to the new supernode to build the node list. After receiving the information, the supernode will send a message to the managed ordinary nodes to inform them that they are managed by the new supernode. The ordinary nodes will be based on This message updates the super node information in its own node list, and the server and these ordinary nodes will communicate with the new super node in the future.

The main difference between a node’s abnormal departure and normal departure is the discovery of the node’s departure. The processing after the node’s departure is found to be the same as that after the normal departure node directly notifies the super node or the server, so here only the abnormal departure of the node is found. to describe.

Abnormal departure of ordinary nodes, abnormal departure of ordinary nodes will be found in two cases. In the present invention, ordinary nodes should regularly send their own survival information to super nodes, and super nodes save the last time of sending survival information of each ordinary node managed by themselves, and super nodes will check this time regularly. If a node has not sent survival information for a long time, it is considered that the node has left the network. Another situation is that the partner node continuously sends data requests to a certain node but does not receive a response, and considers that the node has left the network. At this time, it will notify the super node that the node has left the message, and the super node will detect if there is no response. The node has left the network.

The discovery of abnormal departure of supernodes is somewhat similar to that of ordinary nodes. Supernodes will periodically send their own survival information to the server, so that the server can discover the departure of supernodes. Supernodes, like ordinary nodes, communicate with partner nodes. For data exchange, its partners will also find that the director of the super node has left. In addition, ordinary nodes will dynamically send a request to update the partner list to the super node, and if the super node does not respond, it will send a message to the server that the super node has left the network. The server will update accordingly.

Through the above processing, dynamic information in the network can be discovered in time and effectively updated, so that the network can be dynamically adapted to changes, so that the network is less affected by the withdrawal of nodes.

Therefore, the present invention enables data to be transmitted between similar common nodes on the underlying physical network, reducing the number of hops and time delays in the data forwarding process; The burden reduces the bandwidth consumption; at the same time, the present invention separates node management from data transmission, making maintenance easier.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention without limitation. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be Modifications or equivalent replacements of the technical solutions without departing from the spirit and scope of the technical solutions of the present invention shall be covered by the scope of the claims of the present invention.

Claims (8)

1, a kind of P2P live broadcasting method based on super node, it is characterized in that comprising: The node sends the node joining request information of the desired program to the server, and the server goes to the corresponding program list to match the super node according to the program that the node wants to watch in the node joining request information; The process of matching the supernode is: comparing the flag value of the node with the flag value of the supernode in the server list, if they are the same, the node and the supernode are in the same physical network or are physically close, Complete the match; The server returns the information of the successfully matched super node to the node, and the node sends the node joining request information to the super node according to the information of the super node, and the super node responds to the node joining request Information to add the node as a normal node to the live streaming network, and return the information of the managed normal node to the node; and After the node successfully joins the live streaming network, the node chooses to establish a partnership with other nodes that are physically close to itself, and then transmits data. 2. The P2P live broadcast method based on super nodes according to claim 1, further comprising: the node withdraws from the live streaming network. 3. The super node-based P2P live broadcast method according to claim 1, wherein the process for the node to choose to establish a partnership with other nodes that are physically close to itself is specifically as follows: The node sends a partnership establishment request to the super node to which it belongs; The super node returns its own partner information according to the partnership establishment request; After the node receives the partner information returned by the super node, it selects a node that is physically close to itself and requests to be its partner; If the requested node judges that there is enough bandwidth and the load does not exceed the limit value, it will establish a partnership with the requesting node. 4. The P2P live broadcast method based on supernodes according to claim 3, characterized in that between the node sending a partnership establishment request to the supernode and the supernode returning its own partner information according to the partnership establishment request include: When the supernode receives the request to establish a partnership relationship, it determines that the load does not exceed the limit value after establishing a partnership relationship with the node, and the available bandwidth meets the requirements of the node, and then establishes a partnership relationship with the node. 5. The P2P live broadcast method based on super nodes according to claim 3 or 4, characterized in that after the requested node judges that there is enough bandwidth and the load does not exceed the limit value, after establishing a partnership with the requesting node, it also includes the following step: The requesting node judges whether the number of its partners is sufficient to meet its own needs, and if the number of partners is still relatively small, the requesting node sends a request to establish a partnership with a node with a small delay between itself and establishes a partnership list. 6. The P2P live broadcast method based on super nodes according to claim 5, wherein the process of the node withdrawing from the live streaming network comprises the following steps: The node sends a leaving message to its partner and the super node to which it belongs, notifying the partner and the super node that it will leave the live streaming network; The partner deletes the information of the node from the partner list after receiving the leaving message, and the super node deletes the information of the node from the managed node list after receiving the leaving message, and simultaneously notifies Other normal nodes under management in order to update their node lists. 7. The supernode-based P2P live broadcast method according to claim 1 or 2, wherein the process of matching supernodes further comprises the following steps: If the matching of the super node is not completed, the server adds the node as a super node to the super node list, and notifies the node to join the live streaming network as a super node; and The node joins the live streaming network as a super node. 8. The P2P live broadcast method based on super nodes according to claim 7, characterized in that the process for the nodes to establish partnerships with other nodes that are physically close to them is specifically as follows: When the node acts as a super node, the node will add the nodes with larger available bandwidth and better performance among the ordinary nodes it manages to its partner list, and establish a partnership with it.

Images