Experience the fun of distributed programming - [1]

zhaozj2021-02-16 55

Experience the distributed programming

Redstar81 at 9/9/2003 ~ 13/9/2003 81_redstar81@163.com Tomhornson.student@www.sina.com.cn

First modification time: 13/9/2003 personal text: http://www.9cbs.net/develop/author/netAuthor/redstar81/

Content index

PART1. Summary

I. Introduction to Distributed Technology

Second. Introduction to several excellent distributed computing platforms

Three. Review of several distributed object computing technology

Four. Recommendations of several distributed languages

PART2. Practice

V. Distributed Programming Practice

Objectspace Voyager

2. DCOM

3. CORBA

4. WebServices

5. Sun RMI vs. DOTNET Remoting

6. Linda and Sun Javaspace, IBM TSPACE

7. Jini

8. Emerald vs. dejay

9. Pjama

10. IBM Mobile Computing Interface: Aglets

6. Comprehensive application analysis

1. Calculate technology to simulate high performance parallel computing with multi-threaded and distributed objects

2. Calculate geometric fractal graphs using Vdejay

statement:

1. I am not specializing in the distribution, completely written down this article. I will help distributed computing applicants. 2. Welcome to the level, welcome to correct. 3. Since this paper tells more content, many tells that it is impossible to be too meticulous. And taking into account the problem of the reader group, the theoretical content is omitted as much as possible. If the reader is interested, you can refer to the list of reference books and network resources.

4. Arbitrary reproduced, but please indicate the source; not available for commercial purposes.

PART1. Summary

First, distributed technology overview

Undoubtedly, distribution and parallel theory and technology have long become one of the grasshoids of computer science. In recent years, more and more researchers are involved in it. In terms of distribution, based on, a large number of new theories and technologies have emerged, familiar Known, mobile and intelligent calculations: AS / MAS /, distributed object computing model: DCOM / CORBA / RMI / Remoting, etc. (Basic, these object models are the basis of building computing platforms). Distributed Computing Platform: CORBA / JINI / WebServices /, etc. Object space (tupace space) technology: JavaSpace / Tspace (also used to support the construction of distributed computing platforms). Distributed Language: Emerald / Dejay, etc. / Active network technology, etc.

(Note: As for the above, distributed computing platform, distributed computing object technology, subject space technology, etc., is entirely due to personal knowledge. Please ask some people to correct. Such as DCOM, some books are called distributed computing platforms Of course, you can't say it wrong, but I personally think that it is reasonable to classify the distribution object computing technology. Because it is better than other Corba, Jini is a huge service system based on distributed object technology. MS COM causing Into the distribution computing platform. Then this COM I think it is not only distributed, more concerned about enterprise-class resource coordination / management / system operation performance, so there is no special introduction to COM .)

For 20 years, the distribution calculation has experienced two development routes: the ideal technical route and reality technology route. Ideally technical route concept: deploy distributed operating system, fully managed computers in the system, presented to user single system view In the 1980s, the academic community generally pursued this goal, although many technical results and test systems were produced, they were not accepted by users and markets. Realistic technical route concept: deploy distributed computing environment on the network (middleware Deployment, this may be MS's DCOM and system integration, and provide development tools and public services, support distributed applications, realize resource sharing and collaborative work. Currently, people said distributed computing technology It is usually referred to as a distributed application system development, deployment, management, and maintenance of resource sharing and collaboration as the main application target. Table 1 Three stages of the development of distributed computing technology (invoking from the distribution computing environment>)

First generation

(The 1980s ~ 90s)

Second generation

(the 90s)

Third Generation

(After 2000)

Main problem

Information Sharing

Application interoperability in heterogeneous environments

Intelligent collaborative work

Architecture

Classic C / S calculation model

Object-oriented multi-layer C / S model

Independent Multi-Agent Model

Key technical features

Apply traditional computing concept

And design (such as process call

And files)

Apply object-oriented technology to distributed calculations

Agent's interactive interactive environment

Outcome

Ability to provide a wealth of distributed system services, good distributed system management and typical distributed applications

It has become a core technology that establishes a structure and soft construction standard.

The concept verification system is encouraging and has not been widely used in the maturity of collaborative work.

The first generation technology, from the application of the popular file server, etc., can also see some end from 84 years, Sybase first C / S system database system application and development model. About the third generation, belonging to the field of professional research Things, I have time to introduce you and analyze some classic Agent libraries that can be programmed.

Here we mainly tell the second generation: in fact, the current large-scale application is basically in the second generation, this is, in general, mature applications are often far behind the truth. From the architecture to the basics The idea of programming technology / related technology has no exception (personal feelings? J). Overall, the second generation of core is to organize all resources distributed on the network to object.

So, where is the superiority of distribution techniques? 1. Collaboration based on any network structure. 2. Parallel processing can improve performance. (PS: There is a problem in parallel concepts and distribution concepts. In general, parallel calculations are often first distributed, in fact, we can combine multithreaded technology and distribution computing techniques to simulate parallel computing, to improve computing power) .

1. Have a high degree of scalability. 4. Have a high degree of scalability. 5. Improve the cost performance ratio of the system by resource sharing.

to be continued: