CISC 437 Quiz 2

users, applications, concurrent, multi-threaded

Sources of contention for resources:

• Multiple ________ of an application or transaction executing at one time, e.g. several ATM customers do a withdrawal simultaneously - will compete for the same resources (CPU, hardware, network, etc.)

• Multiple ________________ or systems executing on the same hardware resources at one time

• The application under consideration can have separate ______________ processes

• The application may be _______________ (all threads sharing same space will create contention delay) to handle concurrent requests for different external purposes

system execution model

The _________________________ characterizes the software’s performance in the presence of dynamic factors, such as other work loads or multiple users

system execution model

The _____________________________ aims to solve the contention for resources

software execution model

If the ___________________________ indicates that there are no problems, then you are ready to construct and solve the system execution model to account for contention efforts

system execution model

Benefits of ________________________

• More precise metrics that account for resource contention

• Sensitivity of performance metrics to variations in workload composition

• Scalability of the hardware and software to meet future demands

• Effect of new software on service level objectives of other systems

• Identification of bottleneck resources

• Comparative data on performance improvement options (for new software, calculate new performance metrics and compare with baseline)

server, queue

The system execution model represents the key computer system resources as queues and servers

• A __________ represents a component of the environment that provides some service to the software (e.g. a processor, a disk, or a network element)

• A _________ represents jobs (any requests required for computation) waiting for service (e.g. when a request arrives, if the server is busy, the job making the request must wait until the server is free)

queue, server, wait time, service time, residence time

residence time, utilization, throughput, queue length

Performance metrics of interest for each server are

• ________________ RT: the average time (not a dynamic time that changes per job) jobs spend in the server, in service and waiting

• _______________ U: the average percentage of the time the server is busy, should be <= 1 (e.g. if U = 200, we know it should be reduced to fit workload)

• _______________ X: the average rate at which jobs complete service

• ________________ N: the average numbers of jobs at the server (receiving service and waiting)

performance

The value of ________________ metrics depend on

• The number of jobs (e.g. more jobs = high utilization - more busy)

• The amount of service they need

• The time required for the server to process individual jobs

• The policy used to select the next job from the queue (e.g. the first-come-first-served or priority scheduling)

T, C, B

Calculation of performance metrics:

• ____ = total period

• ____ = completed

• ____ = busy time

B/T

Calculation of performance metrics

• Utilization U = ______

C/T

Calculation of performance metrics

• Throughput X = ______

B/C

Calculation of performance metrics

• Mean service time S = ______

time

Calculation of performance metrics

• Area under graph W = sum of ______ for the number of jobs (add up all of the shaded area under graph)

W/C

Calculation of performance metrics

• Residence time RT = _______

W/T

Calculation of performance metrics

• Queue length N = _______

U

Calculation of performance metrics

• Variable for utilization

X

Calculation of performance metrics

• Variable for throughput

S

Calculation of performance metrics

• Variable for mean service time

W

Calculation of performance metrics

• Variable for area under graph

RT

Calculation of performance metrics

• Variable for residence time

N

Calculation of performance metrics

• Variable for queue length

workload intensity

Solving the queuing model:

• ____________________ is a measure of the number of requests made by a workload in a given time interval

service requirements

Solving the queuing model:

• __________________________ are the amount of time that the workload requires from each of the devices in the processing facility

throughput, arrival rate, jobs-flow balance

Solving the queuing model:

• Assume that the system is fast enough to handle the arrivals, and thus the completion rate or _______________ equals the ________________

• This property is called ____________________

XS, XS/m

Utilization on multiple resources

• U_total = ______

• U_i = _________

N = XR

Little’s law is defined as ___________

throughput, residence time, average

Little’s law:

• Queue length = ______________ x ___________________

• The specifications and results from the simple model are __________ values

• The metrics for a specific job may differ from the average

• The difference depends on the distribution of arrivals and service requirements