Abstract—Queuing network techniques are effective for evaluating the performance of computer systems. We considered a queuing network with two input sources, one is a finite input source and the other is an open input source. We call this as hybrid input source. In the finite input source, the finite number of terminals exists and a job is dedicated to its own terminal. After a think-time at the terminal, the job moves to the server, acquires a part of the memory, and executes CPU and Input / Output (I/O) processing. When the job completes at the CPU and I/O processing, it releases the memory and goes back to its own terminal. On the other hand, in the open input source, the job arrives at the server randomly from outside, acquires a part of the memory, and executes CPU and I/O processing, and goes back to the outside after releasing the memory. However, the queuing network model with memory resource has no product form solution and cannot calculate the exact solutions.
We proposed an approximate queuing network technique to calculate the performance measures of computer systems with hybrid input source in which multiple types of jobs exist. This technique involves dividing the queuing network into two levels; one is "inner level" in which a job executes CPU and I/O processing, and the other is "outer level" that includes termin-als and communication lines. By dividing the network into two levels, we can prevent the number of states of the network from increasing and approximately calculate the performance meas-ures of the network. We evaluated the proposed approximation technique by using numerical experiments and a Monte Carlo simulation, and clarified the characteristics of the system re-sponse time and the accuracy of the approximation.
Index Terms—Performance evaluation, queuing network, central server model, finite input source, open input source.
taru Koike, Hikari Yoshii, Nozomi Miyamoto, Mayuko Hirose and To-shiyuki Kinoshita are with School of Computer Science, Tokyo University of Technology, Hachioji Tokyo, 192-0982, Japan (e-mail: amiro-su.hokuto@gmail.com, c01145528e @edu.teu.ac.jp, c0114508f8@edu.teu. ac.jp, t.kisaragi3so@gmail.com, knoshi@stf.teu.ac. jp).
Cite: Itaru Koike, Hikari Yoshii, Hozomi Miyamoto, Mayuko Hirose, and Toshiyuki Kinoshita, "Approximate Performance Evaluation Method of Computer Systems with Hybrid Input Source," International Journal of Machine Learning and Computing vol. 8, no. 6, pp. 600-605, 2018.