Browse Prior Art Database

Replacement Sector Caching for Optical Disk Drives

IP.com Disclosure Number: IPCOM000113339D
Original Publication Date: 1994-Aug-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 4 page(s) / 134K

Publishing Venue

IBM

Related People

Bish, JE: AUTHOR [+2]

Abstract

A method for reducing the performance impact of reading replacement sectors is disclosed. The proposed method is summarized as a cache memory buffer mechanism dedicated for the storage of some number of replacement sectors. The cache mechanism would allow the drive to avoid the costly seeks associated with reading replacement sectors when a cache 'hit' occurs.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 49% of the total text.

Replacement Sector Caching for Optical Disk Drives

      A method for reducing the performance impact of reading
replacement sectors is disclosed.  The proposed method is summarized
as a cache memory buffer mechanism dedicated for the storage of some
number of replacement sectors.  The cache mechanism would allow the
drive to avoid the costly seeks associated with reading replacement
sectors when a cache 'hit' occurs.

      Optical disk drives allow a fixed number of spare locations per
media surface (both Write-Once Read Many (WORM) and Rewritable) for
defective sector reassignment.

      The Linear Replacement Algorithm is used to handle defective
sectors found after certification.  After this Linear Replacement has
occurred, all reads and writes to that sector occur at the
replacement location (referred to as the replacement sector).  If the
initiator is executing a multi-sector read or write, the drive is
required to seek between the nominal contiguous sectors and any
replacement sectors.  These added seeks cause a considerable
performance impact since the average seek time to and from one
replacement sector is roughly equivalent to the time it takes the
drive to read approximately 170,000 bytes of data (or about one
hundred seventy 1024-byte sectors).

      The proposed method is summarized as a cache memory buffer
mechanism dedicated for the storage of some number of replacement
sectors.  The cache mechanism would allow the drive to avoid the
costly seeks associated with reading replacement sectors when a cache
'hit' occurs.

The mechanism sequence follows:

1.  Upon loading of the media, the drive will initialize an empty
    memory buffer dedicated for replacement sector caching (this
    cache can be a subset of available memory if the memory is also
    used for read-ahead and/or other data caching applications).  The
    drive will also initialize an empty table to be used for
    maintaining pointers to the buffer.

2.  During drive idle periods after initialization of the replacement
    sector cache and between handling initiator commands, the drive
    will read replacement sectors from the media, storing each one
    sequentially in the replacement sector cache along with an entry
    in the pointer table indicating the original defective sector
    location (for look-up) and the cache memory location (Fig. 1).

3.  Any asynchronous event (initiator command, eject request, etc.)
    will be allowed to interrupt the current replacement sector
    caching process such that all cached replacement sectors just
    prior to the current replacement sector in cache will remain
    valid and available for use.

4.  After the asynchronous event handling has completed, the
    replacement sector caching process will resume where it left off
    until the dedicated replacement sector buffer is full.

5.  When the number of replacement sectors on the media exceeds t...