Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Hit Ratio Control for Dasd Control Unit Performance Testing

IP.com Disclosure Number: IPCOM000103119D
Original Publication Date: 1990-Jun-01
Included in the Prior Art Database: 2005-Mar-17
Document File: 1 page(s) / 61K

Publishing Venue

IBM

Related People

Wang, LT: AUTHOR

Abstract

Performance testing using artificial workloads is very important in the operation of a controller for direct access storage devices (DASD). The application program evaluation tool is a convenient artificial workload generator used for the purpose of testing the performance of the control unit, especially a control unit including cache store devices. Generally, curves are generated showing the relationship between the I/O access rate, the cache store hit ratios, and the I/O response time using a series of operations from the evaluation tool test cases.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 59% of the total text.

Hit Ratio Control for Dasd Control Unit Performance Testing

      Performance testing using artificial workloads is very
important in the operation of a controller for direct access storage
devices (DASD).  The application program evaluation tool is a
convenient artificial workload generator used for the purpose of
testing the performance of the control unit, especially a control
unit including cache store devices. Generally, curves are generated
showing the relationship between the I/O access rate, the cache store
hit ratios, and the I/O response time using a series of operations
from the evaluation tool test cases.

      The method of generating a set of performance curves is started
by creating a relatively large data base from which random selections
will be made whenever there is an I/O access.  The larger times that
the access is made directly from the cache store generally means that
a shorter I/O response time for a given I/O rate.  The data set is
then divided into N records of identical block size wherein N is much
greater than one and is a fixed number.  The data set is then
logically divided into N divided by M groups, each group consisting
of M consecutive records, where M is greater than or equal to one and
less than or equal to N.

      For an I/O access, one record will be randomly selected from
any of the M consecutive records of any of the groups. A minimum I/O
rate is set for the test case using the application program
evaluation tool. ...