МЕТОДИКА ЗАЩИТЫ РАСПРЕДЕЛЕННЫХ ВЫЧИСЛЕНИЙ В МНОГОАГЕНТНОЙ СИСТЕМЕ

The organization and protection of distributed computing based on a multi-agent system for solving problems of multivariate modeling is considered. When modeling, the choice of one of many options may require a search of a huge set of parameters unavailable for a high-speed computer. Distributed computing is used to reduce the time required to solve such problems. There are many different approaches to the organization of distributed computing in a computer network - grid technology, metacomputing (BOINC, PVM and others). The main disadvantage of most existing approaches is that they are designed to create centralized distributed computing systems. Distributed computing is organized on the basis of a multi-agent system on the computing nodes of any computer network. When used as a computing environment a computer network on a large scale can cause threats to the security of distributed computing from the intruders. One of these threats is getting the calculation about the result by the attacker. A false result can lead in the modeling process to the adoption of not optimal or wrong decision. A method of protection of distributed computing based on a multi-agent system from the threat of false results is developed. The assessment of the degree of information security in multi-agent systems with different structure. Comparison between information securities in these two systems is provided.