Home > other >  The sixth chapter P2P principle and practice
The sixth chapter P2P principle and practice

Time:10-11

1. The principle of P 2 P
P2P (Peer - to - Peer) the word if only understood as a network topology, it can't call it the new things, because the Internet long before many years had just was born, and the Internet is a large network, P2P structure between people is completely "point to point" way of communication, now called the Server and the Client does not exist, it can be seen as a peer-to-peer (P2P) the original situation,
P2P technology belongs to Overlay Network (Overlay Network) category, is relative to the client/server (C/S) model is a way of the Network information exchange, in the C/S model, the data distribution using dedicated server, multiple clients are from the server to get data, the advantage of this model is: the consistency of the data is easy to control, it is easy to management system, but the shortcoming of this model is: because the number of servers only one (even if there are multiple also very limited), system vulnerable to single point failure; A single server in the face of numerous clients, due to the ability of CPU, memory size, the limitation of network bandwidth, can the client service is very limited at the same time, poor scalability, P2P technology is to solve these problems and put forward a peer-to-peer network structure, in P2P network, each node can get service from other nodes, can also provide services to other nodes, so that large terminal resources being used up, solve the two disadvantages of C/S mode,
2. The development of P2P can be roughly divided into three stages
The first phase, from 1999 to 2000 years or so, the P2P technology developed by application to the academic research and development,
The second phase, from 2001 to 2003 years or so, with the development of P2P technology, start from academic research to commercial application, many system and standard are also constantly improve,
The third stage, since 2004, the P2P technology from a broad consensus to a more comprehensive, practical, and now the main problem with P2P network has been solved, core mechanism, has formed the overall framework, also formed a consensus on major issues,
3. The characteristics of P2P technology
Decentralization, scalability, robustness, high cost performance, privacy protection, load balance
4. The classification of P2P networks
The development of the P2P network has experienced different age, generally can be divided into four generations, the first generation is typical of the centralized P2P network topology structure; The second generation of typical for whole distributed unstructured P2P network topology structure and the whole distributed structured topology structure; The third generation of typical is mixed a distributed P2P network topology structure, and P2P technology development up to now, people are constantly research can better adapt to the modern application demand of the fourth generation of P2P network topology, it is an important direction in the development of P2P technology,
Among them, the biggest advantages of centralized topology maintenance is simple, high efficiency of resource discovery;
Fully distributed unstructured topology of P2P Network is in overlapping Network (Overlay Network) adopted random graph mode of organization, node degrees obey the Power - law rule (Power law), so that they can quickly find goal node, in the face of the dynamic change of Network embodies the good fault tolerance, so has good availability;
DHT class structure to the dynamic adaptive nodes to join/exit, has a good scalability, robustness, the uniformity of the distribution of node ID and the organization ability
Half from the centralized distributed topology structure and the advantages of the fully distributed unstructured topology, choose high performance (processing, storage, bandwidth, etc) of nodes as a super, super nodes in each store the information of the other parts of the nodes in the system, the algorithm is only between the super node forwarding, super node and then forward query request to the appropriate node leaves, half the distributed structure is a hierarchical structure, constitute a high speed super nodes forward layer, super node and ordinary node constitute a few levels of responsibility, adopt the structure of the most typical case is KaZaa,

  • Related