Browse Prior Art Database

Improved Reclamation Process through Selective Overwriting of Inactive Data on Sequential Access Media

IP.com Disclosure Number: IPCOM000029462D
Original Publication Date: 2004-Jun-29
Included in the Prior Art Database: 2004-Jun-29
Document File: 2 page(s) / 31K

Publishing Venue

IBM

Abstract

Sequential access media and devices, like tape, store data in a sequential order. In other words, the files are stored on the medium in the order they are written. This means when new files are written to a sequential access media, they are either written from the beginning or they are appended at the end of the last file. If a sequential medium is written from the beginning, it is considered a new medium, and all data files which reside on the tape from previous operations are overwritten. Appending data is possible from a certain position until the end of the medium is reached, and only the data that exist previous to the beginning of the write location are preserved. It is not possible to insert data in the middle of a sequential access medium. Eventually, some of these data files are no longer active and they become obsolete. To free up space on a media, previously, all active files from a piece of media were then migrated to a new piece of media, regardless of their position on the media, as the means of eliminating the obsolete files. The process of freeing up space on a media is also called reclamation. This publication offers a solution which minimizes the migration of active files onto new pieces of media.

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 39% of the total text.

Page 1 of 2

Improved Reclamation Process through Selective Overwriting of Inactive Data on Sequential Access Media

Sequential access media and devices, like tape, store data in a sequential order. In other words, the files are stored on the medium in the order they are written. This means when new files are written to a sequential access media, they are either written from the beginning or they are appended at the end of the last file. If a sequential medium is written from the beginning, it is considered a new medium, and all data files which reside on the tape from previous operations are overwritten. Appending data is possible from a certain position until the end of the medium is reached, and only the data that exist previous to the beginning of the write location are preserved. It is not possible to insert data in the middle of a sequential access medium. Eventually, some of these data files are no longer active and they become obsolete. To free up space on a media, previously, all active files from a piece of media were then migrated to a new piece of media, regardless of their position on the media, as the means of eliminating the obsolete files. The process of freeing up space on a media is also called reclamation. This publication offers a solution which minimizes the migration of active files onto new pieces of media.

The solution which minimizes the migration of useful files onto new pieces of media has at its basis a preferred measure of effectiveness based on performance. Performance for reclaiming space is defined by the amount of data to be copied off the media to gain more space for the storage of active data. The performance is also dependent on the time it takes the device to seek to where the data resides on the sequential access media.

Via a flowchart, Figure 1shows the preferred method by which the storage management system can select the most efficient alternative for reclamation. In order to perform space reclamation, the storage management system first selects a medium for reclamation in step 70. Step 70 may be executed at a fixed increment of time, such as monthly or annually, by storage management software. Alternately step 70 may be executed on demand by the system administrator. The process then flows to step 71, where the partitions on the medium are identified and a calculation is made of the amount of active data stored in each partition. For simplicity in Figure 1, we only consider three equal-capacity partitions: beginning, middle, and end.

The process then flows to decision step 72, where the determination is made whether there is active data is in the beginning partition of the sequential media. If the decision is yes in step 72, the process flows to decision step 73, to see if there is active data in the middle partition of the sequential media. If the decision is no in step 73, the process flows to step 77, where the determination is made whether there is active data in the end partition of the sequential...