Extendable system structure-based distributed system and application method thereof

Abstract

The invention provides an extendable system structure-based distributed system and an application method thereof. The separation of configuration of distributed application and a distributed structure is achieved through the separation and independent operation of configuration servers; furthermore, under the condition of requiring extending the distributed structure, a distributed system is not required to be modified and recoded greatly, the waste of resource replication is reduced, the configurability of the distributed system structure is achieved furthest, and the advantages of the distributed system can be fully utilized. Through the configuration of the configuration servers, the separation of data right, namely data access and data write, is achieved; through the setting of data access attributes of various meta-information servers, a space after the extending of the distributed structure can be fully utilized, the access pressure of the servers is reduced, the efficiency of anetwork and a bandwidth is effectively utilized, and the performance of the whole system is improved.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a distributed system based on an extensible architecture and an application method of the system. Background technique [0002] The so-called distributed is relative to centralization, and a distributed system is a collection of independent computers, and these computers look like a computer to system users; almost all large-scale software systems are bound to be distributed, and a distributed The largest and most familiar distributed system is the set of computers, software, and services that make up the World Wide Web. [0003] The traditional distributed system realizes the shunting of data, and its distributed structure is as follows: figure 1 As shown in , the meta-information server stores data information and defines various information operations, and the load server is responsible for receiving client operations, adjusting the access balance between meta-information s...

AI Technical Summary

Problems solved by technology

[0005] In addition, distributed systems usually solidify the distributed structure in applications, that is, distributed systems and distributed applications are highly coupled; since different distributed applications have different distributed structure requirements, when distributed applications require When changing, it is often necessary to re-develop and deploy the distributed structure; that is, most of the server information is solidified in the system program. After the system has been running for a period of time, it is often necessary to increase the number of meta-information servers as the amount of information in the system increases. It is necessary to modify the original system program and add the information of the new meta-information server in the program; regardless of the traditional distributed system or the model proposed by the above method, there is a big gap in the scalability of the distributed system. insufficient:

[0006] First, every time a new meta information server is added to a traditional distributed system, the source program of the system needs to be modified, compiled and then re-run, which easily causes a lot of waste of resources;

[0007] Second, when the meta-information server is expanded in the distributed system using the above-mentioned prior art method, because the original meta-information server has formed a vector ring structure, and there is a certain degree of data between adjacent vector rings. Therefore, when a new meta information server is added to the distributed system, it is necessary to break the ring structure and rebuild a new vector ring, and it is necessary to revisit the data stored in the meta information server when rerunning To establish redundancy; therefore, this method makes the coupling between metadata servers more tightly, and the time and resources consumed in server expansion in distributed systems are also greater;

[0008] For this, there is also a server system and method based on network storage and scalable architecture (application number: 02103274.2) in the prior art. The server is connected, when receiving the request from the client but unable to find the information needed by the client, the meta-information server submits the request information to the load server for analysis and judgment, and finds out the meta-information server containing the specified information; because the system The load server in the model has a certain judgment ability, so although it can support some simple extensions, its defects are quite obvious:

[0009] First, although the total capacity of the distributed system expands after the expansion of the meta-information server, the dynamic adjustment of the existing meta-information server cannot be realized because the separation of data reading and writing has not been realized, so the existing As far as the meta information server is concerned, the benefits of the expansion of the total capacity cannot be fully utilized; this makes it impossible to effectively use the expanded meta information server in the process of increasing information;

[0010] Second, in this model, user requests are submitted to the meta-information server through the Internet, so that user applications can directly access the total source of information on the client side, which reduces the security of information to a certain extent and is easy to provide for various information through the Internet. The network provides access to the system's attacks, which increases the operating risk of the system to a certain extent.

Images