Efficiency Analysis of a Simulation Model of Cloud Computing in the Case of Multiple Server Failures

Nowadays, the use of distributed (cloud) resources becomes more and more widespread. Despite the high reliability of equipment used in call centres, failures do occur. Whereas long-term storage systems have high reliability due to multiple duplication, the failure of a server on which a task is executed in a specialized computing system (SCS) can cause the failure of the entire system. To avoid it, a SCS simulation model built in the Simulink (SimEvent) environment, which is classified in the queuing theory as G/G/n/∞, is investigated. It is characterized by multiple input flows with an infinite queue and has two feedback loops, which reflect the reprocessing situations in the case of a server failure or absence of a solution at the first service attempt. The system architecture is focused on the parallel execution of a certain class of tasks, divided into data-independent subtasks. The main operational criterion for such a system is the completion of all arrived tasks in a time not exceeding the deadline of their execution. The purpose of this work is to investigate the reliability of the system for various probabilities of server failures. Since the performance and efficiency of the system have already been studied in our previous works, the main focus here is on the impact of server failures on the overall productivity and efficiency under various laws of input flows.