Return to search

An Automated Multi-agent Framework For Testing Distributed System

Testing is a part of the software development life cycle (SDLC) which ensures the quality and efficiency of the software. It gives confident to the developers about the system by early detecting faults of the system. Therefore, it is considered as one of the most important part of the SDLC. Unfortunately, testing is often neglected by the developers mainly because of the time and cost of the testing process. Testing involves lots of manpower, specially for a large system, such as distributed system. On the other hand, it is more common to have bugs in a large system than a small centralized system and therefore there is no alternative of testing to find and fix the bugs. The situation gets worst if the developer follows one of the most powerful development process called continuous integration process. This is because developers need to write the test cases in each cycle of the continuous integration process which increase the development time drastically. As a result, testing often neglected for large systems. This is an alarming situation because distributed system is one of the most popular and widely accepted system in both industries and academia. Therefore, this is one of the highly pressured areas where lot of developers engaged to provide distributed software solutions. If these systems delivered to the users untested, there is a high possibility that we will end up with a lot of buggy systems every year. There are also a very few number of testing framework exist in the market for testing distributed system compared to the number of testing framework exists for traditional system. The main reason behind this is, testing a distributed system is far difficult and complex process compares to test a centralized system. Most common technique to test a centralized system is to test the middleware which might not be the case for distributed system. Unlike the traditional system, distributed system can be resides in multiple location of different corners of the world. Therefore, testing and verification of distributed systems are difficult. In addition to this, distributed systems have some basic properties such as fault tolerance, availability, concurrency, responsiveness, security, etc. which makes the testing process more complex and difficult. This research proposed a multi-agent based testing framework for distributed system where multiple agent communicate with each other and accomplish the whole testing process for a distributed system. The bullet proof idea of testing centralizes system has been reused partially to design the framework so that developers will be more comfortable to use the framework. The research also focused on the automation of testing process which will reduce the time and cost of the whole testing process and relief the developer from re-generating the same test cases over and over before each release of the application. This paper briefly described the architecture of the framework and communication process between multiple agents.

Identiferoai:union.ndltd.org:siu.edu/oai:opensiuc.lib.siu.edu:theses-2118
Date01 May 2013
CreatorsHaque, Ehsanul
PublisherOpenSIUC
Source SetsSouthern Illinois University Carbondale
Detected LanguageEnglish
Typetext
Formatapplication/pdf
SourceTheses

Page generated in 0.0019 seconds