Browse Prior Art Database

A method to eliminate long read I/Os

IP.com Disclosure Number: IPCOM000127533D
Publication Date: 2005-Aug-31
Document File: 1 page(s) / 6K

Publishing Venue

The IP.com Prior Art Database

Abstract

This article descrives about cache management for the storage devices. Response time for read command can be reduced by shrinking pre-fetch operations.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 56% of the total text.

Page 1 of 1

A method to eliminate long read I/Os

  When read command issues, HDD performs pre-fetch (i.e. zero latency read and Lookahead read) to improve their benchmark performance. These operation are also done even if cached write commands are exist.

Previous desigen If LBA overlap is detected between write cached data and pre-fetch pre-fetch data, read operation should wait for destaging of write cache data. This wait causes long I/O time.

New design If LBA overlap is detected for zero latency read, then shorten the zero latency read to start from the end of the LBA overlap. If LBA overlap is detected for lookahead read, then stop the lookahead read just before the LBA overlap

                |=======|                                                           |=========|                  overlap write data          request range                            overlap write data      |--------------------------------------|+++++++++++++|--------------------------------------------------------------
---|     normal                                                               ...