[Gllug] [Fwd: Server]

[Mike Brodbelt]

On Sun, 2003-11-23 at 19:27, Martin A. Brooks wrote:
> At 16:00 23/11/2003 +0000, you wrote:
> >http://hardware.devchannel.org/hardwarechannel/03/10/20/1953249.shtml
> 
> This was Slashdotted and ridiculed appropriately.  It's not scientific, 
> it's not even relevant because the tests were poorly thought out - can 
> anyone say "file fragmentation"?

If you read my post, that's why I said it wasn't scientific..... The
follow on debate was what was interesting, not the article. Several
viewpoints advanced in the debate mentioned areas where the drives were
physicaly different (different head motors, etc). My experience suggests
it's more than just electronics. If you just sit next to a load of SCSI
disks, the head clatter is enormous compared to IDE. That's a physical
difference, for a start.

> The difference between SCSI and IDE is nearly purely electronics and 
> interface presentation. Some manufacturers ship better quality, i.e. with 
> less defects, platters with SCSI than they do with IDE, some don't.

I've seen that viewpoint advanced a lot. I've never seen compelling
evidence that it represents the true state of affairs. Why do you
believe that it's accurate?

> I routinely use both IDE and SCSI in mission critical systems, the choice 
> for SCSI usually revolves around the number of disks that are going to be 
> hooked up to the system. That said, 3ware's utterly stunning RAID cards 
> soon makes this point moot.

The original poster's complaint was that SCSI systems cost more. 3ware
recommend the 8-port card for workgroup servers, which is quoted as over
500 euros, so goes quite some way towards addressing the price
difference.

Mike.


-- 
Gllug mailing list  -  Gllug at linux.co.uk
http://list.ftech.net/mailman/listinfo/gllug