Browse Prior Art Database

Sequential Staging Control

IP.com Disclosure Number: IPCOM000047940D
Original Publication Date: 1983-Dec-01
Included in the Prior Art Database: 2005-Feb-08
Document File: 3 page(s) / 42K

Publishing Venue

IBM

Related People

Benhase, MT: AUTHOR [+2]

Abstract

A cached storage hierarchy has a control for limiting sequential data in the cache portion. Normally, record N plus the next record (N + 1) is kept in the cache. The record N - 1, the record referred to just prior to record N, is destaged from the cache. In the event record N - 1 or other previously sequentially processed records are referred to, that record is staged as a non-sequential record subject to normal replacement control. If the record N + 1 is replaced, even though sequential, it will not be staged again before a reference to it from a user. A host is attached to the storage hierarchy through a cache. The cache has a usual directory for locating data. The storage hierarchy is completed by a backing store, such as a main memory, direct-access storage device (DASD), tape recorder, and the like.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 3

Sequential Staging Control

A cached storage hierarchy has a control for limiting sequential data in the cache portion. Normally, record N plus the next record (N + 1) is kept in the cache. The record N - 1, the record referred to just prior to record N, is destaged from the cache. In the event record N - 1 or other previously sequentially processed records are referred to, that record is staged as a non-sequential record subject to normal replacement control. If the record N + 1 is replaced, even though sequential, it will not be staged again before a reference to it from a user. A host is attached to the storage hierarchy through a cache. The cache has a usual directory for locating data. The storage hierarchy is completed by a backing store, such as a main memory, direct-access storage device (DASD), tape recorder, and the like. The term record used in this article may be a host- defined record, a track or portion of a track on a DASD, a fixed block of data such as 4KB, a predetermined sized block from a tape recorder, and the like. Such a storage hierarchy can be called a cached DASD subsystem when DASD is employed. Such a subsystem provides fast access cache for the DASD backing store which combines the better attributes of the cache (high-speed access) with the better attributes of the backing store (low cost). The performance objective is to maximize the number of host accesses which can be resolved by accessing a cache copy of the data rather than by accessing the backing store. When sequential data is indicated, there is a remote possibility that the data occupancy of the cache can be principally the sequential data. Such sequential data, when once or twice referenced by the host, may not be referenced again for a long time; hence, the data content of the cache does not necessarily provide the optimum data content. It is desired to limit the contents of the data cache for sequential data sets, such as taught in [*]. While that store technique involved a detector for detecting sequential data, enhanced operation can also be provided by receiving an intent signal from the host that sequential data is being processed for a given portion of the backing store, such as a given DASD. In an I/O system, the sequential access intent signal indicates that references made by the current command chain are part of a physical sequential reference pattern throughout multiple command chains. There are three elements to sequential reference control, the sequential stage ahead (stage N + 1), sequential limiting (destage N - 1), and inhibiting staging of all writing made to the backing store in the sequential mode. The inhibiting of a s...