Q1) Let networks illustrated in the figure below. Suppose computers in institutional network send out 13 requests per second to "origin servers" illustrated in figure. As each request is very small, we suppose there is no delay in sending out these requests and they take 1 second to arrive at "orgin servers" and web response data take another 1 second to come back at Internet-side router on "access link". Each web response object average size is 100kbits.
Now congestion could occur at access link when web response objects come back (we suppose there is no delay in institutional network LAN). Using M/M/1 queue to model access delay in the 1.5Mbps access link. That is to say, average response time (it is the time for a web response object arrives at access link until it enters into institutional LAN) is E[T]=1/(m-n), where n is arrival rate of objects to access link and m is service rate of the access link.
a) Determine total average response time when no institutional cache is used. Response time is time from browser sends out request until it receives web response object. (Hint: total delay includes Internet delay and access link delay)
b) Now assume institutional cache is used. Hit rate for cache is 0.6. Determine the total average response time.