Browse Prior Art Database

Automatic Optical Volume Selection in a Mass Optical Storage Device

IP.com Disclosure Number: IPCOM000037269D
Original Publication Date: 1989-Dec-01
Included in the Prior Art Database: 2005-Jan-29
Document File: 2 page(s) / 14K

Publishing Venue

IBM

Related People

Freeman, HL: AUTHOR [+2]

Abstract

Described is a technique of dynamically assigning optical volumes to a hierarchical directory identifier (called 'path') when that path is first encountered. A list of all the volumes in the mass optical storage device is searched to find a candidate for containing the data objects associated with the given path. The search is based on several parameters that are used to maintain high performance of a mass optical storage device (MOSD).

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 Optical Volume Selection in a Mass Optical Storage Device

Described is a technique of dynamically assigning optical volumes to a hierarchical directory identifier (called 'path') when that path is first encountered. A list of all the volumes in the mass optical storage device is searched to find a candidate for containing the data objects associated with the given path. The search is based on several parameters that are used to maintain high performance of a mass optical storage device (MOSD).

A MOSD may contain any number of optical platters. These platters are entered into the device and initialized by a system operator. In addition, an application environment exists that is accessing data on these optical platters. It is difficult from an application perspective to predict when a volume will fill up and when a new volume should be used. It is easier for the application to provide a new path for an object. The described technique addresses this problem by selecting optical volumes dynamically from a pool of volumes that are pre- initialized and entered into a MOSD.

This technique starts with empty platters in the MOSD and first writes to the 'A' side of these platters. The 'A' side is determined when the platter is first inserted into the MOSD. By writing to one side of the platters first, the probability is fairly low that concurrent accesses will occur to both sides of a platter. This is due to the normal decrease in accesses to data as the data ages.

The system operator controls how often a complete optical platter becomes full by entering a number of empty platters in the MOSD. If the MOSD is filled to capacity with empty platters, then all volumes on the 'A' side are written before any of the 'B' side volumes. However, if the operator places only a finite number of empty volumes in the MOSD, then this technique switches to use the 'B' side as soon as all the 'A' side volumes are filled. This permits both sides of an optical platter to fill up at a faster rate. The latter technique may be completely acceptable from a performance perspective if...