Browse Prior Art Database

Storage Compression ALGORITHM

IP.com Disclosure Number: IPCOM000062278D
Original Publication Date: 1986-Nov-01
Included in the Prior Art Database: 2005-Mar-09
Document File: 2 page(s) / 14K

Publishing Venue

IBM

Related People

Mundy, P: AUTHOR [+2]

Abstract

This article describes a class of algorithms that may be used by a storage manager in a transaction processing system to control the storage utilization of the system. This class includes the program compression algorithm used by CICS/VS Releases up to and including Release 1.6.1, but it also includes algorithms that have a much less severe effect on response times at high transaction rates. An example of such an algorithm is given. In a high-performance transaction processing system such as CICS/VS several resource managers compete for the use of virtual storage. For example, the program manager loads the programs required by the transactions currently being executed.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 1 of 2

Storage Compression ALGORITHM

This article describes a class of algorithms that may be used by a storage manager in a transaction processing system to control the storage utilization of the system. This class includes the program compression algorithm used by CICS/VS Releases up to and including Release 1.6.1, but it also includes algorithms that have a much less severe effect on response times at high transaction rates. An example of such an algorithm is given. In a high- performance transaction processing system such as CICS/VS several resource managers compete for the use of virtual storage. For example, the program manager loads the programs required by the transactions currently being executed. When a program is no longer required, the program manager does not delete the program but leaves it in virtual storage in the expectation that it will be required again by the same or another transaction. Eventually the storage will become full of such programs such that the storage utilization will become critical. Time will then have to be spent to retrieve storage by deleting those programs that are no longer in use. Typically this will cause a stop-go situation: during a busy period storage utilization will increase until it becomes critical. Program compression will then be invoked and any transactions in the system will experience a delay until the program manager has finished its scan and removed the inactive programs. After program compression many of the programs just deleted will again be required, causing further delays to individual transactions while they are reloaded. The aim of this invention is to avoid the erratic response times in the situation described above by a more progressive reduction in the over-use of storage as peak transaction rates are approached. THE ALGORITHM The Storage Manager As virtual storage is allocated and released, the storage manager maintains a record of the ratio of the storage currently in use to the total virtual storage available. When this ratio rises above a threshold n1%, the storage manager will call each resource manager to notify it of a target for its storage usage. The target thus notified will be a function of the overall storage utilization. As storage utilization increases further, the resource managers are notified of successively higher targets until, at an overall storage utilization ratio of n2%, the storage manager calls the resource managers to notify them of a target of 100%. If storage utilization falls the resource managers will be notified of lower targets until at utilization n1% the targets are back to 0%. In practice, the calculation of the storage in use will be done on the basis of the numb...