RELIABILITY ANALYSIS OF RING, AGENT AND CLUSTER BASED DISTRIBUTED SYSTEMS

The introduction of pervasive devices and mobile devices has led to immense growth of real time distributed processing. In such context reliability of the computing environment is very important. Reliability is the probability that the devices, links, processes, programs and files work efficiently for the specified period of time and in the specified condition. Distributed systems are available as conventional ring networks, clusters and agent based systems. Reliability of such systems is focused. These networks are heterogeneous and scalable in nature. There are several factors, which are to be considered for reliability estimation. These include the application related factors like algorithms, data-set sizes, memory usage pattern, input-output, communication patterns, task granularity and load-balancing. It also includes the hardware related factors like processor architecture, memory hierarchy, input-output configuration and network. The software related factors concerning reliability are operating systems, compiler, communication protocols, libraries and preprocessor performance. In estimating the reliability of a system, the performance estimation is an important aspect. Reliability analysis is approached using probability.