A Survey on Queuing Systems with Parallel Servingof Customers. Part II

This paper is a continuation of the survey of the “fork-join” queuing systems (in the westernclassification) or the systems with splitting of queries. Interest in such systems is explainedby a wide range of problems that can be solved with their help, since in fact it is a matter ofparallel processing of data and their applications. For example, this may concern the analysis ofdisk arrays, cloud computing, high-performance services and even the process of picking ordersin a warehouse. In the first part of the survey, the main features of the described model (andrelated systems) and its construction were introduced. Also the detailed description of theapproach to obtaining an accurate expression of the average response time in the case of twodevices was presented as well as several methods of approximate analysis of this characteristic(the case when the number of devices is more than two). This part of the survey is devotedto the description of other existing methods for approximating the average response time. Inparticular, the approaches of the approximate analysis of the response time are as follows: thematrix-geometric method, the analysis with the help of order statistics for various types ofdistribution of the service time of subqueries.