本站原创
Abstract. Distributed computing system is a science by which a complex task that need for large amount of computation can be divided into all pieces and calculated by more than one computer， and we can get the final result according to results from each computer. This paper considers a distributed computing system running in the all and medium—sized computer clusters to solve the problem that single computer has a low efficiency， and improve the efficiency of large—scale computing. The experiments show that the system can effectively improve the efficiency and it is a viable program.
Keywords： Distributed computing； Web Service； Dynamic Link Library
Distributed processing technology is not new， and a lot of Internet—based distributed projects are running. Processing capacity of some of the larger distributed computing projects can meet or exceed the world's fastest supercomputers. This article will explore a simple design of all and medium—sized cluster—based distributed computing systems.
Web Service is based on the standard interface [4]， so they can communicate even they are written by different languages and running on different operating systems. Therefore， Web Service is a good way to implement distributed applications for the different systems. Web Service uses standard protocols （e.g. HTTP） to exchange XML messages to communicate with the client and a variety of resources.
3. Development of Distributed Computing System
The fundamental problem that Web Services technology need solve is using the standard Internet protocol， providing an intermediate layer for building distributed multi—layer application system which is Internet—based， cross—platform， cross—language and cross—hardware， so that between every two applications they he the characteristics of loose and coupling， component—oriented and cross—platform， thus we can achieve the integration of heterogeneous application systems and provide a strong technical support for service enterprises. This chapter designs a Web Services—based distributed computing system in accordance with the general principles and core technology of distributed system design.The distributed computing system is based on the structure of C/S （client / server）， and composed of management server and client systems that he installed software agent （Figure 1）. The distributed computing system works as follows： When a computer joins in the ranks of the distributed computing， it would download and install the client software to the local to be one of the client systems of the distributed computing project. The server is responsible for dividing large computational problem into several aller computing tasks. When the client system is idle， it sends a request to the server； the server accepts the request and assigns all computing tasks to the client system. The intermediate results are first stored in the local database before the completion of task computation， and after ending the task the final results are sent to the server.
Fig. 1： Structure of the distributed computing system
First， the system is achieved by using C# language and Web Service on .NET platform， and it ses information and the calculation results of all tasks in the database of server. The basic workflow of this distributed system is：
1） The client Host communicates with server and request the GetTask（） method to get a description of the program to be executed. This description includes the name and GUID （Global Unique Identifiers） of the task program， and it is sent to the client in form of XML. If the task description is empty， it means the server—side doesn’t he executable tasks.
2） The client sends a request to the server based on the GUID of the task to obtain the task program， and then the server sends task program in Zip to the client. The client unzips it and ses local path of the task program. The client can call the task program and resources by the path of task.
3） After getting the task program， the client continues to reque
4） In the process of processing data， the client needs to send report to the server at intervals to report the stage of task completion， and the server accepts the report as a log.
5） Finally， when it completes the task， the client integrates the data processing results and sends them to the server side to be stored in a database by the server.The client program for each task contains several function blocks， such as setting the logging， setting the path of task program， the initialization method， data processing block， and returning the results.
As a rapidly developing technology， the development prospect of distributed computing is very optimistic because of its advantages of low operating costs and high computational performance. The dynamic distributed computing system for all and medium—siz
References
ZHAN Ji—xiang： Research on Distributed Application System Based on Web Services （2008）
Information on http：//baike.baidu.com/view/437
[4]HU Min， FU Li： Comparison about Several Typical Distributed Computing Technology. Computer Knowledge and Technology Vol.6，No.5，February 2010， pp.1244—1246
[5]MA Xiao—yan： Research of Distributed Application System based on the Web Services （2006）
[6] Information on http：//zh.wikipedia.org/zh—cn/XML
Keywords： Distributed computing； Web Service； Dynamic Link Library
1. Introduction
Today， the PC has begun to enter millions of households， more and more computers are idle， and the potential of the CPU cannot fully be exploited even a computer is turned on. The emergence of Internet makes it to be possible that we can use all the idle computers. Distributed computingcollects the idle resources on the Internet， divides problems which we need an enormous computing power to solve into many aller parts， then assigns these parts to many idle computers for processing， and combines all these calculation results to get the final results. If we say a particular job is distributed， then the number of computer involved in this work源于：毕业设计论文致谢www.udooo.com
is definitely not one， but a computer network. It is obvious that distributed computing has strong data processing ability.Distributed processing technology is not new， and a lot of Internet—based distributed projects are running. Processing capacity of some of the larger distributed computing projects can meet or exceed the world's fastest supercomputers. This article will explore a simple design of all and medium—sized cluster—based distributed computing systems.
2. Related Technologies
2.1 Dynamic Link Library
DLL （Dynamic Link Library）is a library containing the code and data that can be used by multiple programs at the same time. DLL provides a way to make the process can call the function that does not belong to the executable code of it. The executable code of a function is contained in a DLL， the DLL contains one or more functions which he been compiled， linked and stored separately with the process using them. Multiple applications can simultaneously access a single copy of the DLL in memory. In our distributed computing system， task programs that the server sends to the client are tranitted in the form of a DLL.2.2 Web Service
Web Service [3] is the core of the new generation of distributed systems. It is a service—oriented architecture technology which provides services through standard Web protocols to ensure the applications of different platforms can be interoperable. According to the W3C， Web Service is a software application identified by URL， its interface and the binding can use XML documents to define， describe and discover， the messaging is based on Internet Protocol， and it uses XML to support direct interaction with other software applications.Web Service is based on the standard interface [4]， so they can communicate even they are written by different languages and running on different operating systems. Therefore， Web Service is a good way to implement distributed applications for the different systems. Web Service uses standard protocols （e.g. HTTP） to exchange XML messages to communicate with the client and a variety of resources.
2.3 .NET Platform
.NET [5] is a development platform provided by Microsoft to meet the needs of the development of Internet， and it’s particularly suitable for network programming and web services. Because of standard communication protocols in a .NET environment， the communication between applications on different platforms is ailable. .NET platform includes the .NET framework and .NET development tools. .NET framework is the basis of the entire development platform. .NET development tools include VisualStudio.NET integrated development environment and .NET programming languages. VisualStudio.NET integrated development environment is used to develop and test applications， and .NET programming languages include VB， VC + + and C # which were used to create applications.2.4 XML
XML [6] is a standard published by the W3C in 1998， and it has become a Web standard to represent and tranit data on the Internet. The use of XML in the .NET framework is very common， including in the configuration files， the C # source code comments， and Web services. In Web services， XML is a data representation protocol， and as the representation of message in the SOAP protocol it is also used to describe the protocol itself. In this distributed systems， all messages are passed源于：论文库www.udooo.com
in the form of XML.3. Development of Distributed Computing System
The fundamental problem that Web Services technology need solve is using the standard Internet protocol， providing an intermediate layer for building distributed multi—layer application system which is Internet—based， cross—platform， cross—language and cross—hardware， so that between every two applications they he the characteristics of loose and coupling， component—oriented and cross—platform， thus we can achieve the integration of heterogeneous application systems and provide a strong technical support for service enterprises. This chapter designs a Web Services—based distributed computing system in accordance with the general principles and core technology of distributed system design.The distributed computing system is based on the structure of C/S （client / server）， and composed of management server and client systems that he installed software agent （Figure 1）. The distributed computing system works as follows： When a computer joins in the ranks of the distributed computing， it would download and install the client software to the local to be one of the client systems of the distributed computing project. The server is responsible for dividing large computational problem into several aller computing tasks. When the client system is idle， it sends a request to the server； the server accepts the request and assigns all computing tasks to the client system. The intermediate results are first stored in the local database before the completion of task computation， and after ending the task the final results are sent to the server.
Fig. 1： Structure of the distributed computing system
First， the system is achieved by using C# language and Web Service on .NET platform， and it ses information and the calculation results of all tasks in the database of server. The basic workflow of this distributed system is：
1） The client Host communicates with server and request the GetTask（） method to get a description of the program to be executed. This description includes the name and GUID （Global Unique Identifiers） of the task program， and it is sent to the client in form of XML. If the task description is empty， it means the server—side doesn’t he executable tasks.
2） The client sends a request to the server based on the GUID of the task to obtain the task program， and then the server sends task program in Zip to the client. The client unzips it and ses local path of the task program. The client can call the task program and resources by the path of task.
3） After getting the task program， the client continues to reque
摘自：学术论文翻译www.udooo.com
st for the data to be processed according to the GUID of thetask. If the server does not he data to be processed， then send 0. The client ses all data in the queue to start processing， and each data processing result is also stored in the queue. When all the data are processed， the client sends final results to the server side.4） In the process of processing data， the client needs to send report to the server at intervals to report the stage of task completion， and the server accepts the report as a log.
5） Finally， when it completes the task， the client integrates the data processing results and sends them to the server side to be stored in a database by the server.The client program for each task contains several function blocks， such as setting the logging， setting the path of task program， the initialization method， data processing block， and returning the results.
4. Conclusions
On the basis of in—depth study on distributed computing related technology and based on the reality of distributed computing， this paper analyses the structure and characteristics of distributed system and discusses the technical advantages of Web Service architecture. We propose a simple and cost—effective distributed computing system solution based on the general principles and core technologies of distributed system design. The experiments show that the relevant techniques and methods in the implementation of this system are practicable， and these methods he certain theoretical value and the actual reference.As a rapidly developing technology， the development prospect of distributed computing is very optimistic because of its advantages of low operating costs and high computational performance. The dynamic distributed computing system for all and medium—siz
论文大全www.udooo.com
ed computer clusters we introduce is suitable for general all—scale research organizations. The system can significantly improve the computational efficiency and it is a cost—effective distributed program.References
ZHAN Ji—xiang： Research on Distributed Application System Based on Web Services （2008）
Information on http：//baike.baidu.com/view/437
3.htm
[3]MA Wei： Analysis on Client— server Computing’s some Technology. Computer Knowledge and Technology No.20 （2008）， p.283[4]HU Min， FU Li： Comparison about Several Typical Distributed Computing Technology. Computer Knowledge and Technology Vol.6，No.5，February 2010， pp.1244—1246
[5]MA Xiao—yan： Research of Distributed Application System based on the Web Services （2006）
[6] Information on http：//zh.wikipedia.org/zh—cn/XML