Multiple methods for searching for File Download sources

In P2P research, the main technical problems are: Node Discovery and communication strategies; 2. File upload and download policies; whether a file can be downloaded is an important indicator of user satisfaction. Then, how to discover the file source becomes the object of study in this article.

1) centralized requests: when users discover file sources, the first generation of P2P software is also the most commonly used and most effective method: obtain the address information and file information of the file source from the centralized file directory server (the server filters out online users). Then, the client sends a request to the target through the above information to establish a connection, finally, download is achieved. However, this method has a major defect: 1) if the server crashes, the user will not be able to obtain the file source, and will not be able to use its functions; 2) the user will continuously request the server, as a result, the server is overloaded. The more P2P users, the faster the download.

(2) friend community extension: Maze adopts the concept of friend in maze. It adopts the Six-degree Space theoretical model, and thinks that interaction can be achieved between two people only with six people at most, this is based on the concept of Friends community, that is to say, the clustering of human life determines the predictability of a certain social space. 1) When downloading a file, due to the clustering characteristics of human society, we can easily find that there are many similar resources in the same region, such as the same school and department, therefore, by sending a friend, we can send a search request to the friend's SSN neighbor and easily obtain multiple file sources, providing raw materials for Multi-Point download. 2) Similarly, people are composed of groups. Due to space limitations, they cannot understand other interests and common interests. The server analyzes the downloading behaviors of users, A person's character is wrinkled. Therefore, we can recommend related users and download resources to them, providing users with a new way of mutual understanding. Through the concept of friend community, we can achieve a small range of serverless use even if the client does not log on to the server, according to the local friends information recorded in the past.

3) super node Extension: If a super node (a strong client source with an independent IP Address can store friends in a region, it can be viewed as a small server ), the entire network consists of one or more servers, and the super node to the server interacts with other node information, so as to achieve complete P2P in the region, and the node information can be interacted through a controlled flood broadcast outside the region, this is also the advantage of hybrid P2P. For more information, see eMule.

Problems to be Solved: 1) is the super node feature a small server in real life? How does it interact with other super nodes and what information does it interact? How is the super node scope determined, and how is the communication between the Protocol nodes within its scope?