Question

A system consists of 1 CPU, 1 Disk, and 1 Network Interface. A web server running on this system consists of a process that waits for an HTTP request. Once the request is received, the process services it by fetching the requested file from disk (an I/O operation) and then by sending the file content to the client. Assume that serving a request consists of the following phases:

i. Process uses the CPU for 2 msec (parse request) // CPU is busy here

ii. Process blocks for the Disk for 9 msec (fetch the file) // Disk is busy here

iii. Process uses the CPU for 3 msec (prepare for network I/O) // CPU is busy here

iv. Process blocks for the network for 5 msec (send file to the client) // Network is busy here

v. Process uses the CPU for 1 msec (cleanup) // CPU is busy here

In the above, when a resource is "busy" serving a process, it means that the resource cannot be used by any other process; any other process that needs that resource must block until the resource is available. Assuming that there can only be one process in the system (i.e., a batch processing system, with a multiprogramming level MPL = 1), and assuming that there is always requests for the server to service (i.e., once the process is done with a request, it can immediately start on the next) answer the following questions:

(1) What are the utilizations of the CPU, disk, and network?

(2) What is the capacity (in terms of requests per second) of the web server when MPL=1?

With one process in the system, when a resource is busy, all other resources will be idle. Thus, to improve resource utilization, you decided to use N=2 processes to serve more than one request at a time. In other words, you decided to go for MPL = 2 (i.e., the web server has two processes, each of which serving one of the requests coming into the system). Again, assuming that there are plenty of requests for each process to service, and that the system has been running for a long time (i.e., it reached some steady state) answer the following questions:

(3) For MPL=2 what is the utilization of the CPU, disk, and network?

(4) What is the capacity of the web server when MPL=2

(5) What is the smallest MPL value that will result in one of the resources becoming the bottleneck?

(6) What is the maximum achievable capacity of the web server?

Answer 1