Browse Prior Art Database

Defragmentation improvement on RAID disk device

IP.com Disclosure Number: IPCOM000011893D
Original Publication Date: 2003-Mar-24
Included in the Prior Art Database: 2003-Mar-24
Document File: 3 page(s) / 33K

Publishing Venue

IBM

Abstract

Disclosed is a idea for shortening the processing time of defragmentation for RAID disk devices except RAID-0/1.Conventionally, defragmentation is realized by moving data to optimal position in disk device.In non-RAID and RAID-0/1 disk devices, as one of the effects of defragmentation, I/O response improves.It is because data become close each other physically.However, in RAID disk devices except RAID-0/1, the improvement of I/O response is not supposed to exist.It is because RAID mechanism has managed the physical position on a disk device, and consequently, it is because data don't necessarilly become close each other physically.On this idea, it is enabled to carry out defragmentation without moving data.It is realized by newly introducing a data position managed table on RAID disk devices.Consequently, the processing time of defragmentation can be shortened sharply.From here, "RAID disk device" points out RAID except RAID-0/1.

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 56% of the total text.

Page 1 of 3

Defragmentation improvement on RAID disk device

   Figure 1 is the functional block diagram of this idea. The portion of a dotted line is a new portion of this idea.

"The data position manage table" is newly introduced in RAID disk devices.This table relates data positions of OS view with those of physical disk devices. In this figure, if one reads data on Track-Y on OS view, one gets "Data-A", becouse RAID disk device provides OS with "Data-A" on "Physical position-A" via "The data position manage table". Moreover, the command that updates this "The data position manage table" is introduced newly.When a defragmentation process executed on OS, the I/O function of OS issues this new command to the RAID disk device.This command includes "The position before move" and "The position after move".

Figure 2 shows how "the data position manage table" is updated by the time series.
(1)RAID disk device receives the command that requests to update "the data position manage table".

In this figure, the command requests to move the data on Track-Y to Track-X.
(2)The physical position field of entry "Track-X", which is the position after move, is updated by that of entry "Track-Y".
(3)The physical position field of entry "Track-Y", which is the position before move, is updated by "no data".

Figure 3 shows the status viewed from OS and from RAID disk devices. As a result of updating "the data position manage table", "Data-A" is moved from Track-Y to Track-x in OS view

Figure 1: functional block diagram

RAID disk device

data position manage table

Track-1 Physical-1

... ...

Physical disk drive

... ...

... ...

Track-Y Physical-A

position    contents

Track-X no data

Physical-A

Data-A

New command that updates the data position manage table

 position before move

OS

defragmentation process

Track-X

no data

position   contents

logical disk from OS view

Track-Y

Data-A

 position after move

Figure 2: command processing chart

before update1

the command that requests to update "the data position manage table"

after update

data position manage table

data position manage table

data position manage table

Track-1

Physical-1

Track-1

Physical-1

Track-1

Physical-1

...

...

...

...

...

...

Track-X

no data

position before move = Track-Y

Track-X

Physical-A

Track-X

Physical-A

...

...

...

...

...

...

Track-Y

Physical-A

position after move = Track-X

Track-Y

Physical-A

Track-Y

no data

...

...

...

...

...

...


.

1

[This page contains 7 pictures or other non-text objects]

Page 2 of 3

Figure 3: The status viewed from OS and from RAID disk devices

logical disk from OS view

position  ...