Browse Prior Art Database

Automatic Initiation of Disk Library Packing

IP.com Disclosure Number: IPCOM000074169D
Original Publication Date: 1971-Mar-01
Included in the Prior Art Database: 2005-Feb-23
Document File: 2 page(s) / 14K

Publishing Venue

IBM

Related People

Henry, GG: AUTHOR

Abstract

This algorithm is used to determine when to physically pack a disk library. The problem of whether to and when to pack a file library occurs on all systems supporting a file library resident in one contiguous area. The problem occurs when a file is "deleted". This operation usually consists of making the file logically nonexistent by deleting an entry in the library directory. The physical disk space occupied by the file still exists, however, and is a "hole" in the library area. If such an unused "null" area is left in the library, a situation may arise in which there is no one contiguous null area large enough to hold a new file to be saved, but the total unused space in the fragmented library is large enough for the file.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

Automatic Initiation of Disk Library Packing

This algorithm is used to determine when to physically pack a disk library. The problem of whether to and when to pack a file library occurs on all systems supporting a file library resident in one contiguous area. The problem occurs when a file is "deleted". This operation usually consists of making the file logically nonexistent by deleting an entry in the library directory. The physical disk space occupied by the file still exists, however, and is a "hole" in the library area. If such an unused "null" area is left in the library, a situation may arise in which there is no one contiguous null area large enough to hold a new file to be saved, but the total unused space in the fragmented library is large enough for the file. Thus, either the files must be physically "packed" together to create a larger contiguous null area, or the new file cannot be saved. Another problem with leaving unused gaps in the disk area for the library is that it potentially increases seek distance to access files in the library and thus reduces performance of data management functions.

Two prior approaches to handling library fragmentation are: (1) to physically pack the library following each file deletion; and (2) to pack the library only upon explicit specification by the user. The first technique has a serious disadvantage in the large amount of time required to physically pack the library, especially on a machine with a relatively slow disk. An advantage is that the user is never concerned with fragmentation of the library. At any point he will get "maximum" performance in referencing files and maximum "usable" free space in the library. The second technique has the advantage of providing a fast file deletion function (only a library directory entry is changed). The disadvantage is that if the user is not aware of physical library fragmentation occurring, he may unknowingly be losing performance and function by not invoking a "Library Pack". It is possible to be "out of library space" as far as the user is concerned even though the library space is almost completely unused.

The technique disclosed here is designed to "automatically" pack the library as few times as possible yes still provide maximum usable library space and adequate performance. The term "automatic" means that the system, not the user, initiates the pack operation. The general determination algorithm for when to pack the file library is to never pack it except under the following conditions.: 1) A new file is to be stored in the library and there is not a contiguous unused area large enough to hold it, and the total of the unused areas in the library is large enough to hold the file. 2) The to...