Research And Design Of P2P-based Online Games

Online games industry is an emerging sunrise industries, has experienced the early stages of the formation of the late 20th century period, and the rapid development in recent years. China's online game industry is in the developing period and quick to mature stage. Online games developing from scratch in China's economic development process, is currently an important part of the Internet economy in China.With the development of online games, the problems of the game mainly based on traditional C / S structure has become increasingly clear: the cost expensive and increasingly heavy burden on the server, a single point of failure and so on.The rise of P2P technology, making the original server model may take P2P model in some online games instead. So cost-effective use of resources and cost-effective may achieve a major breakthrough, which is very favorable to the current status of low-cost design of the game. However, based on P2P technology, there exists excessive transport stream in online games: the exchange of information through the broadcasting network load will increase, reducing real-time P2P environment, and even the system collapsed. Therefore, this paper conducted a study put forward a preliminary solution address to these issues.Interest Management in China has just started. But in foreign countries, a number of universities and research institutes has put forward interest management-related programs in support of large-scale expansion of the MMOG. By studying foreign language information and making improvements, we have come up with a fairly complete mechanism Combined the regional division map of the MMOG and the interest management together.N-Tree-based game regional division methods divide ASS (game world) into regions of varying sizes. Because the region will be conducted from time to time multicast entity status information, the players exchange information among each other and each event must have at least N2 message transmission times. Intensive players together will result the scope of information transmission reach a peak in a small area. The load of the network is so large that may cause network delay or obstruction. The worst-case scenario of the congestion would be making communication down to the traditional situation. In this way, there is not only an increase in the players' computation nodes but also the occupation of network bandwidth. Therefore, when players exceeding the threshold number of a region. the region will continue to be divided. According to the characteristics of the N-Tree, the region can be dynamically adjusted according to the number of regional groups members, so that the number of players in each region in the balance while ensuring that the majority of players receive useful information. Each coordinator of the region is the team manager and also the multicast group manager. This method can be combined very well with the application layer multicast technology. Through the N-Tree structure, adjacent regions can be quickly calculated, which is a good solution to the issue of regional borders. So that is a relatively simple calculation and maintenance. N-Tree can also enable players nodes effectively move to a new location, add a new group and find the node that is closing. Therefore, Region division is the basis of the predictable management mechanism. If there is no rough filtering by region division methods, predictable interest management mechanism will not be in the feasibility.Predictable interest management mechanism is based on the prediction of physical movement and sending different games messages with the different frequency. The entities make frequent movement in the game and the entities which have been influenced change. We using Aura denotes the region that entities (game players) are interested. PAI denotes the regional scope that entities cycle over a period of time and take charge of the maintenance of the entity in Aura and PAI. The news of the play will be divided into three categories according to whether PAI and Aura of the entities are Interactive: APUM (low-frequency transmission); LPUM (a small number of high-frequency information transmission); OAM (a large number of high-frequency transmission). The players node located in the region that is divided by N-Tree and the buffer of the region will receive periodic messages APUM that includes entities sent by coordinator. The players node where PAI interactive entity located in receive LPUM news . The players node where Aura interactive entity located in receive the physical node OAM message. The use of Predictable Interest management reduces the send of useless information and achieve the purpose of less information transmission . We use the distributed architecture with a central server as the infrastructure of the network . The Model includes a central server, supporting server , client. Central server is mainly used for authentication ,managing the players' account and the managing the resources throughout the game, including the type, quantity. The supporting server Used to help customers that can not communicate directly transmit data. Client establish a connection when validating and withdrawing from the game with the central server, and then communicate. After the completion of communication the server disconnect with the central server. In the State of the game, the game clients adopt distributed communication method. The news of the game directly send to the purposed client.Finally, we based on the JXTA technology, use the concept of group to simulate the realization of the interest domain and achieved the desired results.In this paper, N-Tree-based regional division of the game and predictable interest management mechanism still need to be improved and perfected.