Distributed Parallel Virtual Machine: An Object-Oriented Approach
Nanda Gopal, Vishnu Gopal, Vivek Mohan Thampi · 2004
Programming tasks are increasingly becoming prohibitively computationally intensive. While Moore’s law guarantees an almost exponential increase in processing power by the year, more and more, processor-taxing tasks seem to have an even larger appetite. Macro-computers – behemoth machines with impressive speeds measured in teraflops, and aggregating the processing power of thousands of separate micro-processors, are the traditional answer to this problem of program complexity. However, such controlled application servers are hard to maintain, often requiring a dedicated team of computer engineers to monitor, and often extremely costly. If such systems are improperly designed, failure of a single node in the network could bring down the entire server; such systems therefore, unless severely monitored, do not provide a guarantee of program execution. Another solution which employs the same bank of processors, but which does not rely on a cohesive tie-in and permits greater flexibility, is termed Distributed Computing. Unlike supercomputing systems, a distributed computing system