Consider a transaction processing system that contains a "front-end" processor, which accepts and translate user queries arriving from a large number of terminals and formats the replies, and a "back-end" processor running a database management system. The database files are stored on disks A and B that are connected to the "back-end" processor. Each query entered from a user terminal is processed by the "front-end" and then is sent to the "back-end" for further processing. When the answer is found, it is sent back to the "front-end" processor and then to the user's terminal for display. It was found that a transaction requires in average 100 ms on the front-end processor, 150 ms on the back-end processor, 5 visits to disk A and 8 visits to disk B. The duration of an I/O operation is of 30 ms for each disk and 20 ms for display. The response time for a query is considered from the moment of submitting the query to the front-end until the displaying of the reply.

a)Draw an open Queueing Network model to represent the system

b)What is the system bottleneck?

c)What is the minimum average response time (in milliseconds)?

d)Under what conditions will the minimum response time occur? Light load, heavy load or moderate load?

e)What is the maximum utilization at the Front End CPU?

f)What is the maximum utilization at the Back End CPU?

g)What is the maximum possible average throughput of the system (in queries/millisecond)?

h)What happens if the bottleneck device is replaced by another unit twice as fast?