Uniprocessor Scheduling

1
Uniprocessor Scheduling

• Chapter 9

2
Processor Scheduling

• Processor scheduling determines the assignment of
  processes to be executed by the processor over
  time
• Goals of scheduling
• High processor utilization
• less idle time
• Fast throughput
• large number of processes completed per unit time
• Quick response time
• small elapsed time from the submission of a
  request to the beginning of the response
• Fairness
• Low overhead

3
Classification of Scheduling Activity

• Long-term which process to admit
• Medium-term which process to swap in or out
• Short-term which ready process to execute next

4
Queuing Diagram for Scheduling
5
Long-Term Scheduling

• Determines which programs are admitted to the
  system for processing
• Controls the degree of multiprogramming
• Batch jobs
• Whether to create a new process? If more
  processes are admitted
• better CPU utilization - less likely that all
  processes will be blocked
• each process has smaller fraction of the CPU time
• Which job to admit next? May be based on
  priority, I/O requirements, etc.
• The long term scheduler may attempt to keep a mix
  of processor-bound and I/O-bound processes

6
Long-Term Scheduling II

• Interactive programs
• E.g. user connecting to time-sharing system
• Requests not queued up
• Will accept new process requests until the system
  is saturated
• When the system is full a new connection request
  is refused with an error message

7
Medium-Term Scheduling

• Swapping decisions based on the need to manage
  multiprogramming
• On systems with no virtual memory, the size of
  the process is also a criterion
• Done by memory management software and discussed
  in chapters 7 and 8

8
Short-Term Scheduling

• Determines which process is going to execute next
  (also called CPU scheduling)
• The short term scheduler is known as the
  dispatcher
• The scheduler is invoked whenever an event occurs
  that may lead to blocking or preemption of the
  current process. Example events
• clock interrupts
• I/O interrupts
• operating system calls
• signals

9
Short-Term Scheduling Criteria

• Allocate processor time to optimize certain
  aspects of system behavior such as the following
  (Table 9.2)
• User-oriented criteria perceived by a user or
  process
• response time (interactive jobs), turn-around
  time (batch jobs)
• System-oriented criteria
• processor utilization, throughput
• Performance-oriented (quantitative) readily
  measured
• response time, throughput
• Non-performance-oriented (qualitative)
• predictability, fairness
• Thus, the design of a scheduling policy involves
  compromising among several competing requirements
  and depends on the nature and use of the system

10
Definition of terms

• Turnaround time
• Interval of time between the submission of a
  process and its completion
• Response time
• Interval of time from submission of a request
  until the response begins to be received
• Throughput
• Number of processes completed per unit time
• Processor utilization
• Percentage of time the processor is busy
• Fairness
• In the absence of priorities, processes should be
  treated the same no process should starve

11
Using Priorities

• Implemented by having multiple ready queues to
  represent each level of priority
• Scheduler will always choose a process of higher
  priority over one of lower priority
• Problem Lower-priority may suffer starvation
• Then allow a process to change its priority based
  on its age or execution history
• Our first scheduling algorithms will not make use
  of priorities
• We will then consider other algorithms that use
  dynamic priority mechanisms

12
Characteristics of Scheduling Policies

• The selection function determines which process
  in the ready queue is selected next for execution
• The decision mode specifies the instants in time
  at which the selection function is exercised
• Non-preemptive
• Once a process is in the running state, it will
  continue until it terminates or blocks itself for
  I/O
• Preemptive
• Currently running process may be interrupted and
  moved to the Ready state by the OS
• Allows for better service since any one process
  cannot monopolize the processor for very long
• Greater overhead than non-preemptive ones

13
Example to discuss various scheduling policies
Service Time
Arrival Time
Process
1
0
3
2
2
6
3
4
4
4
6
5
5
8
2
Service time total processor time needed in one
CPU-I/O cycle Or Service time total execution
time required
14
First Come First Served (FCFS)

• Selection function the process that has been
  waiting the longest in the ready queue (hence,
  FCFS) also known as FIFO
• Decision mode non-preemptive
• a process runs until it terminates or blocks
  itself

15
FCFS drawbacks

• A process that does not perform any I/O will
  monopolize the processor
• Favors CPU-bound processes
• I/O-bound processes have to wait until CPU-bound
  process completes
• They may have to wait even when their I/O are
  completed (poor device utilization)
• We could have kept the I/O devices more busy by
  giving a bit more priority to I/O bound processes
• FCFS is not an attractive policy on its own
  combine with priority scheme

16
Round-Robin (Time Slicing)

• Selection function same as FCFS
• Decision mode preemptive
• a process is allowed to run until the time slice
  period (quantum, typically from 10 to 100 ms) has
  expired
• then a clock interrupt occurs and the running
  process is put on the ready queue

17
Time Quantum for Round Robin

• Length of the time slice is the principal design
  issue for round robin
• Very short time quanta should be avoided because
  there is processing overhead involved in handling
  the clock interrupt and dispatching function
• Should be slightly larger than a typical
  interaction time to get good response time

18
Round Robin Critique

• Effective in a time-sharing or transaction
  processing system
• Still favors CPU-bound processes
• A I/O bound process uses the CPU for a time less
  than the time quantum and then is blocked waiting
  for I/O
• A CPU-bound process runs for all its time slice
  and is put back into the ready queue (thus
  getting in front of blocked processes)
• A solution virtual round robin (VRR)
• When I/O is completed, the blocked process is
  moved to an auxiliary queue which has preference
  over the main queue
• A process dispatched from the auxiliary queue
  runs no longer than the basic time quantum minus
  the time spent running since it was selected from
  the ready queue

19
Queuing for Virtual Round Robin
20
Shortest Process Next (SPN)

• Selection function the process with the shortest
  expected processing time
• Decision mode non-preemptive
• Short processes are moved ahead of longer jobs in
  the queue
• We need to estimate the required processing time
  (CPU burst time) for each process

21
Estimating average CPU burst

• Simple averaging
• Ti execution time for ith instance of this
  process (total execution time for batch job
  processor burst time for interactive job)
• Si predicted value for ith instance
• S1 predicted value for first instance not
  calculated

22
Estimating average CPU burst

• Exponential averaging
• Sn1 ?Tn (1 - ?)Sn (0 lt a lt 1)
• Sn1 ?Tn (1 - ?)?Tn-1 (1 - ?)i?Tn-i
  (1 - ?)nS1
• when ? is close to 1, weight is given to most
  recent observations
• quickly reflect rapid change in observed quantity
• when ? is smaller, the averaging is spread over
  several observations
• Eg ? 0.8
• Sn1 0.8Tn 0.16Tn-1 0.032Tn-2 0.0064Tn-3
  
• S1 0 gives greater priority to new processes

23
(No Transcript)
24
S1 0
25
S1 0
26
Shortest Process Next Critique

• Possibility of starvation for longer processes as
  long as there is a steady supply of shorter
  processes
• Lack of preemption is not suited for time sharing
  environments
• CPU bound process gets lower priority (as it
  should) but a process doing no I/O could still
  monopolize the CPU if it is the first one to
  enter the system
• SPN implicitly incorporates priorities shortest
  jobs are given preferences

27
Shortest Remaining Time (SRT)

• Selection function the process with the shortest
  expected remaining processing time
• Decision mode preemptive
• When a new process with a smaller remaining time
  enters the ready queue, the running process is
  preempted for the new process

28
Shortest Remaining Time Critique

• Can be considered a preemptive version of SPN
• Does not have a bias for longer processes
• Risk of starvation of longer processes still
  exists
• As with SPN, need an estimate of processing time
• Also, elapsed service times must be recorded
  adds overhead
• Gives better turnaround time than SPN because
  short jobs are given immediate preference

29
Highest Response Ratio Next (HRRN)

• Selection function the process with largest
  value of response ratio (RR)
• w time spent waiting for the processor
• s expected service time
• Decision mode non-preemptive

30
Highest Response Ratio Next Critique

• Expected service time must be estimated as before
• Shorter jobs are favored because they have
  smaller denominator (larger ratio)
• Age of the process is also considered therefore,
  longer jobs do get a chance to get past competing
  shorter jobs
• waiting time for longer jobs increases, thus RR
  increases and the longer job gets scheduled

31
Multilevel Feedback Scheduling

• Preemptive scheduling (using time slices) with
  dynamic priorities
• Several ready to execute queues with decreasing
  priorities P(RQ0) gt P(RQ1) gt ... gt P(RQn)
• A new process is placed in RQ0
• When it is preempted (finishes its time quantum),
  it is placed in RQ1. Next time it is preempted it
  is placed in RQ2 and so on, until it reaches RQn
• I/O-bound (short) processes will stay in higher
  priority queues. CPU-bound jobs will drift
  downward.
• Dispatcher chooses a process for execution in RQi
  only if RQi-1 to RQ0 are empty, hence longer jobs
  may starve

32
Multiple Feedback Queues

• FCFS is used in each queue except for lowest
  priority queue where Round Robin is used

33
Time Quantum for feedback Scheduling

• With a fixed quantum time, the turnaround time of
  longer processes can stretch out alarmingly
• To compensate we can increase the time quantum
  according to the depth of the queue (RQi 2i)
• Longer processes may still suffer starvation.
  Possible fix promote a process to higher
  priority after some time

34
Algorithm Comparison

• Which one is best?
• The answer depends on various factors such as
• the system workload (extremely variable)
• relative weighting of performance criteria
  (response time, CPU utilization, throughput...)
• Summary of algorithms See
• Table 9.3
• Table 9.5, Figure 9.5