On Multiresource Queues with Multiserver Customers

Queueing systems with multiserver customers anew attracted the attention of researchers due to a number of modern applications such as the supercomputers, storage and wireless transmission. In such a system, each customer requires a random number of servers simultaneously. The models are hardly analytically tractable even in simple cases due to the non-work-conserving property and huge state space. However, it is important for applications to capture more details of the system, such as the shared resources (e.g. memory). In this paper we address this problem by appropriate model modifications.