[KLUG Hardware] Re: hardware RAID cards (was Informal Survey)

[Bryan J. Smith]

--=-EkalOtb2x0cDmS3iKBkC
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Thu, 2002-10-24 at 18:37, Mike Williams wrote:
> At 4-7 clock cycles per read, ouch.

It gets worse for today's CPUs.

A 2GHz processor is 0.5ns latency.

People talk about "full speed" L1 and L2 cache, but that's not exactly
true.  ;-P

Normally the L1 cache is usually same as logic, so no delay.  The L2 is
usually pretty fast, maybe 1-2ns, so 2-4 cycles.

Then we get to main memory.  40-70ns =3D 14-25MHz.  Ouch!  That equates to
80-140 cycle delays!

> How about reads, since a workstation does more read than write?  In=20
> sequential reads I could see the sorting of access requests that SCSI=20
> allows (forgot the technical term) being an advantage.  Is the=20
> controller on an Escalade smart enough to do that sort of thing with IDE?

Yes.  _True_ hardware ATA RAID cards offer this.  And it's the #1 reason
why "Trick BIOS" ATA RAID cards _suck_ -- because they don't.

> My aforementioned ignorance is showing again.  I looked up levels 3 and=20
> 4 and couldn't tell the difference.  They both seem to be RAID 5, but=20
> with all parity data on the same drive.

Correct.

> What's the difference between 3 and 4?

Just the algorithm.  RAID-3 is designed for workstations, RAID-4 is
designed for large file servers.  NetApp has marketed RAID-4 as a
"better solution" for NFS servers.

> The advantage to it being that reads don't have to hit the=20
> parity drive at all, and sequential writes keep the parity data=20
> contiguous, right?

Correct.  But lots of random writes will kill it.  RAID-5 is best for
application/database servers, RAID-4 is better for large file servers.

--=20
Bryan J. Smith, E.I.            Contact Info:  http://thebs.org
A+/i-Net+/Linux+/Network+/Server+ CCNA CIWA CNA SCSA/SCWSE/SCNA
---------------------------------------------------------------
           limit      guilt   =3D     { psychopath,
         remorse->0                    innocent }


--=-EkalOtb2x0cDmS3iKBkC
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQA9uKtYDjEszaVrzmQRAlFSAJ0RXa7PG1oRXBPiB8y/gKQ3wEbKkACfdWy8
CPi/Bg8h7vP6KgxTYxW9DUQ=
=twXZ
-----END PGP SIGNATURE-----

--=-EkalOtb2x0cDmS3iKBkC--