Browse Prior Art Database

Track Overflow in a Relative Byte Address Environment

IP.com Disclosure Number: IPCOM000079950D
Original Publication Date: 1973-Oct-01
Included in the Prior Art Database: 2005-Feb-26
Document File: 4 page(s) / 52K

Publishing Venue

IBM

Related People

Ostrowski, RT: AUTHOR [+2]

Abstract

A device independent approach to paging I/0 page slot selection on direct access devices is shown. It provides a common interface for access method and paging I/O at a logical level above physical addressing. It describes a format layout and selection algorithm which will provide efficient sequential processing within a cylinder. The reduction in rotational delay provides efficient processing of a random group of records within a cylinder.

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 38% of the total text.

Page 1 of 4

Track Overflow in a Relative Byte Address Environment

A device independent approach to paging I/0 page slot selection on direct access devices is shown. It provides a common interface for access method and paging I/O at a logical level above physical addressing. It describes a format layout and selection algorithm which will provide efficient sequential processing within a cylinder. The reduction in rotational delay provides efficient processing of a random group of records within a cylinder.

In an environment where the relative byte address (RBA) oi the data within the data set is used to reference or index the data and physical records (e.g., in VSAM record management) on a direct access device, it is necessary to convert the RBA to a physical device address. One term needs to be defined at this point. An allocation unit is an arbitrary unit of allocation chosen mainly for the convenience of programming support. In order to perform the conversion with some efficiency, certain characteristics of the recording are required as follows: The physical blocks must be all the same size, i.e., fixed-length block recording. There must be the same number of bytes of data recorded on each cylinder. There must be the same number of bytes of data recorded in each allocation unit. There must be the same number of bytes of data recorded on each track.

The following describes the format and method of utilization of direct access, space using track overflow for such applications as paging.

It is necessary to describe what is commonly referred to as the next sequential record problem. On most direct access devices where the data is recorded such that the physical records start at the same relative position on each track (as would be the case with fixed-length blocks and no track overflow), there is not sufficient gap time to allow the processing of one record and then perform a head switch to get to the next sequential record within the same cylinder, but on a different track. Where this can become a performance problem, artificial means are necessary to space the records differently or otherwise overcome the problem.

The use of track overflow will result in the staggering of the starting position of the records on a track. This is one approach to solving the next sequential record problem, which is very crucial in providing efficient paging support in virtual systems. Of course, there is also the obvious advantage of utilizing the available storage space more fully.

After writing as many fixed-length blocks as will fit on a track, there is a certain amount of available space which could be used to record data on the track. This can be utilized to record a partial block using track overflow, if the conditions can be maintained which will allow an RBA to be converted efficiently. This can be done given the following input: P = Physical blocksize; K = Integral number of blocks per track (whole number of blocks which will fit on one track); TC = Numb...