Browse Prior Art Database

Erase Detection and Error Recovery for Optical Files

IP.com Disclosure Number: IPCOM000036958D
Original Publication Date: 1989-Nov-01
Included in the Prior Art Database: 2005-Jan-29
Document File: 6 page(s) / 175K

Publishing Venue

IBM

Related People

Kulakowski, JE: AUTHOR [+4]

Abstract

Rewritable optical disk drives using Magneto-Optic (MO) media have a unique characteristic when compared to magnetic DASD or tape devices. The drives must erase the MO media prior to writing. Since the write operation occurs on a second revolution and the media is removable, the media can be left in the erased state after a power loss situation or if the disk is physically removed prior to completing a write operation.

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 6

Erase Detection and Error Recovery for Optical Files

Rewritable optical disk drives using Magneto-Optic (MO) media have a unique characteristic when compared to magnetic DASD or tape devices. The drives must erase the MO media prior to writing. Since the write operation occurs on a second revolution and the media is removable, the media can be left in the erased state after a power loss situation or if the disk is physically removed prior to completing a write operation.

The error recovery procedures to handle erased sectors on the media are complex. With magnetic media, a sector could be partially written with loss of power or media removal and a future read opera

(Image Omitted)

tion would result in an ECC error. The host would try to recover the data, and, if it was not recoverable, put the sector out of service. At worst, only one sector would be in error even if the command had been a multiple sector write. For MO optical disk products, many sectors could be left in the totally erased state on a multiple-track write operation. If normal error recovery procedures were followed, the error recovery time would be lengthy and alternate sectors required would be excessive, resulting in a disk user capacity loss. A mechanism is needed to discriminate between a media error or a media error induced by a machine/operational error such as a power loss, and allow the error recovery procedure to be performed in a timely manner.

Elements required to solve this problem are a circuit which detects an erased area on a disk, a storage area in a Non-Volatile RAM (NVRAM) for the Volume Serial Number (VOLSER) of the disk in the drive, and a storage area in an NVRAM for the Relative Block Address (RBA) of the current operation.

(Image Omitted)

Fig. 1 shows the possible states of a sector if power is lost or media is removed during a write operation. A sector could be erased, partially erased or partially written, as shown. A fourth state is shown which is a defective sector with a defect, such as a dust particle on the media in the Sync/MD area. If a read is attempted on the sectors as shown, the results are different. For Case 1 (an erased sector) or Case 2 (a partially erased sector), the Phase-Locked Loop (PLL) will not lock on the data since there is no sync area so the sector will not be found. For Case 3 (a partially written sector), a read will result in an ECC error. The problem is that Case 4 (a media defect in the Sync/MD), will result in a sector-not-found condition similar to Case 1 and 2. If the loss of Sync/MD in case 4 were due to a spot of dust on the media, the data could be recovered by cleaning

(Image Omitted)

the disk which is to be part of the error recovery procedures. If the sector was still bad after cleaning, it would be put out of service and an alternate would be assigned.

1

Page 2 of 6

This same procedure would be very time consuming if applied to a large number of erased sectors that could appear due to a power loss af...