We observe the closed system for 30 minutes, during which 1600 tasks are completed, from 12 terminals. Each terminal (source of tasks) has think time of 12s between receiving response to one task, and submitting next task. System has CPU and two disks, one is fast, other is slow. Fast disk can deliver block in 1/2 of time that slow disk would take. During observation period, 32000 accesses happen at fast disk, and 12000 accesses at slow disk. During observation period, CPU is busy for 1080s, fast disk is busy for 400s, and slow disk is busy for 600s.
a) Explain the response time for jobs in observed system?
b) As function of N, number of terminals, give high-load bounds for throughput and response time; also provide low-load bounds.
c) How many terminals can we support, while keeping response time below 3 s?
d) Assume we plan to replace CPU with CPU whose speed is double (therefore time to execute each job would decrease so it becomes 1/2 of current time). Evaluate number of terminals we can support, with response time below 3s. "