SIMULATION MODEL OF REDUNDANT MACHINE-TO-MACHINE EXCHANGE WITH ORGANIZATION OF QUEUES FOR ACCESS TO AGGREGATED CHANNELS

Subject of Research. The paper presents the study of computer systems in providing exchanges through redundant channels in order to analyze the possibilities of reliability increase and timeliness of computer systems interaction as a result of redundant transfers through aggregated channels, taking into account the implementation options in computer nodes of distributed separate queues for access to each channel, or general queue for access to all channels. Method. The method is based on the development of simulation models of the considered options for redundant exchange through aggregated channels. Main Results. The redundant exchange efficiency is shown taking into account the organization of distributed queues in the nodes for access to aggregated channels. For the first option of redundant exchange, separate queues are organized at each computer node for access to each channel. The incoming request is copied k times (depending on the criticality to the timeout), and each copy is placed in one of the queues. For the second option, one common queue is organized for all channels in each computer node, and each incoming packet is entered in this queue (a request for its transmission). When a request is issued from the general queue, k copies of the transmitted packet are generated, and each copy is transmitted through one of the n channels as far as the node is granted access rights to it. A specific feature of the reserved service is the formation of k copies for each request, issued for service in different channels. A service is considered to be successfully completed if at least one of the k created copies of the request (package) is correctly executed in the specified time. The existence of the optimal reservation multiplicity of transmitted copies of packets is shown; the efficiency area of the reserved machine-to-machine exchange is determined. Practical Relevance. The results can be used in the design of highly reliable computer systems, including real-time ones.