|
Home > Archive > Data Storage > May 2004 > code theory behind RAID systems.-
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
code theory behind RAID systems.-
|
|
| Gustavo V?zquez 2004-05-30, 11:11 am |
| Hi all. I'm triying to begin a code theory investigation about the
redundancy codes in RAID systems. I will like to know if there are any
papers about this type of codes and where could I get them.
Thanks in advance.
Gustavo Vázquez.-
| |
| Anton Rang 2004-05-30, 11:11 am |
| gustavov@internet.com.uy (Gustavo V?zquez) writes:
> Hi all. I'm triying to begin a code theory investigation about the
> redundancy codes in RAID systems. I will like to know if there are any
> papers about this type of codes and where could I get them.
Well, almost all commercial RAID systems use simply parity.
RAID 2 uses a Hamming code, along the lines of what's used for RAM.
There are some interesting error codes proposed to tolerate multiple
disk failures, though I'm not sure these are actually in real use.
One of these ("RAID 6") is P+Q redundancy, which uses Reed-Solomon
coding.
There's a nice overview of RAID in the June 1994 issue of ACM
Computing Surveys. If you need to know about the mathematical
structure of the codes, I'd suggest a textbook on error correcting
codes.
-- Anton
| |
| Mark Landin 2004-05-30, 11:11 am |
| On 13 May 2004 11:35:13 -0700, gustavov@internet.com.uy (Gustavo
V?zquez) wrote:
>Hi all. I'm triying to begin a code theory investigation about the
>redundancy codes in RAID systems. I will like to know if there are any
>papers about this type of codes and where could I get them.
Do you mean things like the parity-checking code?
Parity checking is a pretty well-studied concept in computer science.
A properly constructed Google search should provide you with the
fundamentals. The latest-and-greatest algorithms may be proprietary
and a considered a "trade secret" by the companies which develop them;
those may be more difficult to track down.
| |
| Adrian 'Dagurashibanipal' von Bidder 2004-05-30, 11:11 am |
| -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Clinging to sanity, Mark Landin mumbled in his beard:
> Parity checking is a pretty well-studied concept in computer science.=
> A properly constructed Google search should provide you with the
> fundamentals. The latest-and-greatest algorithms may be proprietary
> and a considered a "trade secret" by the companies which develop them=
;
> those may be more difficult to track down.
If not google, I'd try searching the libraries of the ACM and the IEEE =
- I
guess most of the algortihms were published in some paper, and I bet mo=
st
trade secret-type algorithm are just modified published algorithms.
cheers
- -- vbi
- --=20
Jetzt ist der Herr Bush Pr=C3=A4sident, und weil ihm wieder langweilig =
ist,
will er endlich den Saddam loswerden. Der Herr Bush hat n=C3=A4mlich ke=
ine
Praktikantin.
-- http://bush.d0t.de/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: get my key from http://fortytwo.ch/gpg/92082481
iKcEARECAGcFAkCmT0pgGmh0dHA6Ly9mb3J0eXR3
by5jaC9sZWdhbC9ncGcvZW1h
aWwuMjAwMjA4MjI/ dmVyc2lvbj0xLjUmbWQ1c3VtPTVkZmY4NjhkMTE4
NDMyNzYw
NzFiMjVlYjcwMDZkYTNlAAoJEIukMYvlp/fWkJYAoOyCYjufSz8TYANy6QyTeeNv
k2lrAJ4qG3zWXJajZOD6bjDxzqq2XVx6Fw=3D=3D
=3DtXx3
-----END PGP SIGNATURE-----
| |
| Gustavo V?zquez 2004-05-30, 11:11 am |
| Yes I mean so. I know the algorithms but I would like to know the
usage of that codes (for example, in wich disks you put the pariy bit,
the redundancy, etc).
Mark Landin <mark.landin@tdwilliamson.com> wrote in message news:<uak9a0t6dmq8358t8911da3p837qgauo1v@4ax.com>...
> On 13 May 2004 11:35:13 -0700, gustavov@internet.com.uy (Gustavo
> V?zquez) wrote:
>
>
> Do you mean things like the parity-checking code?
>
> Parity checking is a pretty well-studied concept in computer science.
> A properly constructed Google search should provide you with the
> fundamentals. The latest-and-greatest algorithms may be proprietary
> and a considered a "trade secret" by the companies which develop them;
> those may be more difficult to track down.
| |
| Nik Simpson 2004-05-30, 11:11 am |
| Gustavo V?zquez wrote:
> Yes I mean so. I know the algorithms but I would like to know the
> usage of that codes (for example, in wich disks you put the pariy bit,
> the redundancy, etc).
A good place to start is the original Chen & Patterson et al. research
papers from Berkely, two good ones start with are:
http://www.buyya.com/superstorage/chap1.pdf
http://www.csie.fju.edu.tw/~yeh/res...-raid_intro.pdf
As to where the parity information is stored, it's a pretty simple list
RAID-2, IIRC, the Hamming code data is stored on four disks for a set of 10
data disk, five disk for a set of 25, the number goes up with the gumber of
drives in the stripe.
RAID-3 all parity stored on a single disk, i.e. if there are N disks in the
data set, then the parity is stored on the N+1th disk.
RAID-4 same principle as RAID-3
RAID-5 Parity information distributed across all disks in the set, i.e. no
single drive dedicated to parity data.
|
|
|
|
|