14 Scheduling Basics I

14 Scheduling Basics I - CS-350 Fundamentals of Computing...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: CS-350: Fundamentals of Computing Systems Page 1 of 12Lecture Notes © Azer Bestavros. All rights reserved. Reproduction or copying (electronic or otherwise) is expressly forbidden except for students enrolled in CS-350. Resource Management Basics When we discussed the simple M/M/1 system earlier in this class, we assumed that upon completion of service, the server (e.g., CPU or Disk) would simply pick the “next available” request in the queue. Specifically, we assumed that the server would use a FCFS policy. However, in our analysis of the M/M/1, we did not really make use of FCFS in any of our derivations! And, as a matter of fact, when we discussed GPS versus FCFS, we saw that the actual dispatching discipline does not actually affect the average performance of the system (e.g., average server utilization or average response time). The keyword here is “average.” While “on average” the response time for all requests may be the same under various dispatching disciplines, the response time would clearly be different for different “types” of requests. We motivate this by an example: Consider two Poisson arrival processes (A1 and A2), both feeding into a single queue. The resulting system is still an M/M/1 (since the combination of the two Poisson arrival processes A1+A2 is also Poisson). Let’s contrast that M/M/1 system when a FCFS dispatching principle is used versus the same system when a dispatching discipline that gives requests from A1 a higher priority than those in A2 (but uses FCFS for all requests within the same group). Under FCFS, there would be no difference between the turnaround times for A1 processes and for A2 processes (e.g., a process of type A1 or A2 would experience a turnaround time of 10 seconds on average). Under the “priority scheduling” policy, however, we would expect A1 requests to see a better turnaround time than A2 requests (e.g., a process of type A1 may experience a turnaround time of 8 seconds on average whereas a process of type A2 may experience a turnaround time of 12 seconds on average1). Hence, the manner in which we schedule processes will have an impact on the quality of service extended to various types of processes. In this chapter, we look more closely at some basic scheduling (or more generally “resource management”) concepts. When is Resource Management Necessary (or when does it make a difference)? Resource management is needed when a resource is to be shared by competingprocesses of differentcharacteristics, and especially when the resource utilization is expected to be high (i.e., the resource is heavily used). Another way of saying this is that resource management will not make a big difference when any of the following is true: (1)The resource is lightly loaded—if there is not much demand for a resource then it really doesn’t matter how we schedule it!...
View Full Document

This document was uploaded on 04/14/2011.

Page1 / 12

14 Scheduling Basics I - CS-350 Fundamentals of Computing...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online