The implementation of an application-driven functional architecture of a peer-to-peer distributed computing system defined by conceptual and logical models of artificial intelligence. 1. Declarative and logical approaches

Abstract

Relevance and goals. Most network distributed computing systems (DCS) implement “client-server” and “master-slave” applications. It is relevant to create applications with an arbitrary structure consistent with the functional architecture of the DCS. At the same time, the implementation of distributed computing can be significantly facilitated, allowing the use of remote user computers to perform complex data processing. It is shown that the proposed architecture of distributed computing systems is highly flexible and allows the implementation of complex global calculations based on conventional Dijkstra’s and other well-known distributed and parallel programming structures. The purpose of the work is to develop and experimentally test a methodology and technology for implementing distributed computing systems (DCS) on a network of peer-to-peer nodes with a conceptually defined functional architecture. Materials and methods. To ensure interoperability between the functional and system architectures of the RCS, on the one hand, and the implemented application, on the other hand, this work uses an approach based on the logical approach in artificial intelligence, namely, on the logic of first-order predicates with graphical representation of formulas in the form of conceptual graphs. To implement the DCS, a peer-to-peer system network architecture was chosen, which is highly flexible and allows the implementation of complex global calculations based on conventional Dijkstra’s and other well-known distributed and parallel programming structures, on the basis of which the conceptual graph of a distributed application is built. Results. A technique for implementing the functional architecture of a DCS for a given distributed application is proposed. A text file is selected to store the textual representation of the conceptual graph in the application. The application is capable of implementing arbitrary interaction patterns between network nodes. To change the interaction scheme, it is enough to change the configuration file. Conclusions. The application-controlled functional architecture of a peer-topeer distributed computing system, defined by conceptual and logical models of artificial intelligence, is illustrated by a complex software implementation in a real computer network.

Keywords

• distributed computing system
• peer-to-peer application
• functional architecture
• declarative approach
• conceptual models of artificial intelligence
• textual representation of a conceptual graph