Features of distributed computing, taken into account in methods of algorithms optimization for the volume of interprocessor transmissions

The main parameters, considered in methods of algorithms optimization for the volume of interprocessor transmissions

Abstract

In connection with the development of new technologies and the expansion of the use of computers and networks, in particular with the advent of the Internet of things, there is an increasing need to use a combination of computing devices to solve one problem. In this case, most often these devices are geographically distributed and differ from each other by technical characteristics. This article explores the possibility of adapting the methods of parallel computations originally intended for systems with separate memory to distributed systems. The features of distributed systems, ways of representing such systems and algorithms on these systems are considered. The article proposes an approach to the representation of algorithms taking into account the features of a distributed system in the form of a projection of an information graph of the algorithm onto a interconnection graph. This approach allows us to investigate the algorithm in static mode without going through all the network and algorithm states.