Integrated tolerant distributed computing network
The performance of a distributed computing network (DCN) can be determined in relation to various levels of functioning, in particular, when executing com-mands, programs, tasks, and tasks. Here, the performance of the DCN is related to the level of task completion. The task can be completed if two circumstances occur: 1) the task is ready for execution when a certain event occurs for it, for example, related to receiving input data, and 2) there is a free resource, for ex-ample, a computing module (CM) of the network. The impact of task readiness can be taken into account when using the model of the CM functioning as a sys-tem with a dynamic change in the number of tasks [1,2.3]. In this case, each task executed by the processor brings a certain number of tasks to the ready state out of the maximum tasks to be completed. The availability of a free resource un-doubtedly depends on the distribution of tasks across the CM, where they should be performed [4,5,6,7]. In the conditions of degradation of the DCN due to the failure of the CM, it is necessary to provide measures to ensure tolerance so that the degree of perfor-mance reduction is the lowest. there are many methods for ensuring tolerance with circuit and software tools. Next, we will limit ourselves to software tools, the implementation of which is supported by a number of methods. Some methods involve diagnosing the CM using a diagnostic test with a certain period, for ex-ample, when the task is completed by this module. Other methods involve paral-lel execution of the same task by different modules, comparison of the results of their execution with the help of a programmatically implemented decision body, and performing actions based on the results of this comparison [8,9]. Below, we will build an integrated analytical model of the performance of a tolerant computing system based on two models: a model for evaluating perfor-mance based on the number of serviceable CM, and a performance model based on the method of ensuring tolerance.