Working principles of P2P

The evolution of computer networks is swinging between centralization and distribution. In the early days, many users shared large computers, and personal computers were developed from centralized to distributed. Similar situations exist on the Internet. The client (browser)-server method is adopted to use the server in the website collection. Further development will be distributed, and centralized servers will become distributed. Each user terminal is both a client and a server. This is the peer-to-peer (P2P) mode.

In recent years, P2P services on the Internet have developed rapidly and have become the mainstream of broadband Internet services. P2P technology combines users into one network, shares the bandwidth, and processes the information together. Unlike the traditional client-server mode, in P2P mode, each client is both a client and a server. Taking the shared download file as an example, each user of many users who download the same file only needs to download a part of the file and then exchange it with each other. In the end, each user can obtain the complete file.

The first generation of P2P networks adopts the Central Control Network Architecture (1 ). Earlier Napster adopted this structure. It uses a fast search algorithm with a short queue response time. Simple protocols provide high performance and elasticity, but services are easily interrupted.
The second generation of P2P adopts a distributed network architecture (2 ). The central server is no longer used, eliminating problems caused by the central server. Without a central control point, it will not cause all paralysis due to a single fault. It is a real distributed network. Because each search is performed across the network, a large amount of network traffic is generated, resulting in a slow search speed and long queue response time. The performance of your PC and the connection mode determine the network elasticity and performance. This mode has self-organizing (Ad-Hoc) behavior, reducing the cost of the owner and providing scalability. It is particularly suitable for self-organizing (Ad-Hoc) online applications, such as instant messaging.
Third-generation P2P networks adopt a hybrid network architecture (3 ). This mode combines the advantages of the first and second generations, and replaces the central retrieval server with distributed super nodes. The hierarchical quick search improves the search performance and shortens the queue response time. The traffic generated by each queue is lower than that of the second-generation distributed network. The deployment of super smart nodes provides high performance and elasticity. Without a central control point, no fault will cause all paralysis.
Content is distributed and stored in distributed storage devices and customer terminals. The quick retrieval system allows you to quickly find the location of the content distribution and storage. Currently, P2P software is commonly used, including BT, edonky, and Gnutella. These software uses the "fast tracking" technology to form a P2P network and has many advantages that traditional client-server networks do not have. Technically, this not only greatly reduces the time required for file search, but also removes the need for expensive central control hardware devices (servers ). This P2P network uses the processing capability of the terminal's own computer, and the network processing capability increases with the increase of the number of terminal users.

The fourth generation of P2P is currently under development. The main development technologies include dynamic port selection and two-way download. Dynamic port selection: P2P currently uses a fixed port, but some companies have begun to introduce the Protocol to dynamically select the transmission port. Generally, the number of ports is between and. Even P2P streams can be transmitted using the port 80 (25) originally used for HTTP (SMTP) for hiding. This makes it more difficult to identify P2P streams across carrier networks and master their traffic. Two-way download: companies such as Ed and Bt are further developing and introducing two-way stream download. You can download and upload one or more parts of a file in parallel. At present, the traditional architecture requires that the object be uploaded only after full download. This will greatly speed up file delivery.

The above four generations of P2P systems all belong to "unorganized P2P overlapping networks" and have developed rapidly on the Internet. At present, more than half of the bandwidth user traffic is such a P2P stream.

The academia is currently studying "organized P2P overlapping networks ". Including tapestry, Chord, pastry, and can. The new generation of P2P applications that are being studied, including multicast and network storage, all run on this organized P2P overlapping network.

2. P2P applications

Currently, there are three main applications,
(1). P2P content sharing

This includes BT, eDonkey Gnutella, BearShare for search and retrieval, content distribution, network storage, and peercasting for peer-to-peer broadcast. Network Storage gives full play to the ubiquitous advantages of the Internet. Mobile phones, PDAs, laptops, desktops, televisions, various household appliances and sensors can access services through a variety of wired or wireless access networks.

(2) P2P Communication Collaboration includes collaboration, Internet telephone, instant messaging S, and mobile communication

P2P Instant Messaging System (IM) adopts the peer-to-peer connection mode (P2P). The message format uses XML for effective report management, which can provide asynchronous, parallel, reliable, and approximate real-time communication. Supports mobile registration management and mobile instant messaging. Use the IETF standard to ensure intercommunication.

(3). P2P collaborative computing includes collaborative computing, grid, and data content grid.

The basic concepts of grid GRI * and * 2 * collaboration are very similar to those of resource sharing, but different technologies are used. The purpose of grid is to use network resources for large-scale high-performance computing. The network resources used include supercomputer, cluster, specialized devices, and large-scale databases. The grid requires organizations to develop protocols for resource sharing. Security and performance are key issues. Uses authentication and programmable Middleware Based on PKI + X.509 certificates for high-performance computing. P2P cooperation uses network resources such as personal computers and stored digital content to control resources discretely and allow anonymous access. The main advantage is measurability.

Using P2P and GRI * to generate a data content grid is the most attractive and practical method to build a data grid using the ** 2 * technology. A content grid is built based on P2P technology. Data and content in the grid are automatically distributed, allowing users to access recent data.

The video content GRID provides the VOD service for distributed storage.

(4) P2P will usher in a new era of Network Media

In addition to live broadcasting, TV and video programs are recorded and stored on the server in advance. Vod is a working mode, but it occupies too many network resources. P2P provides a new working mode for network TV media. Users can download and store content in their own computers in P2P mode, and then play back and watch the content.

3. P2P requires IPv6

Due to insufficient IP addresses, the IPv4 internet uses a large number of dedicated IP addresses and network address converter Nat. The result is that information about who is the information sender is hidden. P2P requires the ability to identify senders and use the "real-name system" to establish a network society featuring integrity and security. Only IPv6 can achieve this. The establishment of an IPv6 network will provide a secure, reliable, and inexpensive network operating environment for the development of P2P applications.

To run P2P on IPv6 networks, you no longer need to use NAT to control functions and logic, which simplifies the network design. For carriers, there is no need to consider the compatibility between NAT and its software. In the remote end, servers are not required, which can reduce the operational burden.

In short, the operation and cost are lower than when IPv4 and Nat are used.

Applicable to various communication modes: client-server, peer-to-peer P2P, mobile secure instant communication, instant connection, and end-to-end security. The IPsec Policy server provides a policy file to connect to each Party as needed, which is easy to set: you only need to register on the web.

4. Respond to peer-to-peer business challenges

P2P has become a killer application of broadband. P2P applications develop rapidly around IP address audio and video file sharing. P2P applications currently account for 50-60% of bandwidth traffic (daytime) to 90% (evening), and 40% of enterprise users.

How to identify and control P2P streams, how to protect copyrights and stop the flow of pornographic content has become a headache for enterprise network administrators and access operators. Various methods are being studied, such as buffering P2P streams on the operator's network server to reduce outbound traffic.