Richard A. Steenbergen ras at
Fri May 25 23:57:31 UTC 2001

On Fri, May 25, 2001 at 03:11:13PM -0400, Wojtek Zlobicki wrote:
> Last time that I checked, adding a drive to a raid array or NETAPP was
> a great deal more than $99. Lets not count the cost of upgrading a HD
> in a lowly Linux box (not bashing Linux here).  IDE is also not
> sufficient enought in order to sustaint the write rate of that these
> servers demand (let alone the IO).  5200 RPM drivers are not a
> solution for ISPs.

Ack. I promised myself I was going to stay out of this thread but I don't
think I'm going to be able to resist. If you're going to insult IDE you
should at least know what you're insulting it for.

Your average 60gig ($140) or 80gig ($200) 5400rpm IDE Maxtor drive will
have no problems reading or writing at 30MB/s sustained. I have built
software IDE raid setups that have no problem reading at 50MB/s sustained
with boatloads of CPU to spare, and are not light on the seek ability
especially as the number of drives increase. Your average 7200rpm IBM
75gig drive (a bit more expensive, $250/ea) will perform a lot better then
that. I'm looking at an extremely loaded web server right now pushing over
90Mbit/s at over 600 hits/sec and over 400 drive seeks per second, off a
500gig raid-5 built entirely from cheap 7200rpm IDE drives.

That's not your multi-terabyte mail and news server for hundreds of
thousands of users, but I bet you its a lot cheaper. The only area where
IDE really loses is logistics and cable management, if you're going to
keep drives balanced on seperate IDE channels. With UDMA/66 controllers at
$14/ea from Fry's, there is no technical reason you can't get some great
performance out of IDE drives at extremely cheap prices. If you want to
take it a step more towards business, look into 3Ware products for IDE
drives. Don't forget what the I in RAID stands for.

But enough about off-topic stuff, I don't want to distract from the next
line of discussion which will be if you didn't write your email in vi you
shouldn't have the right to send it.

Richard A Steenbergen <ras at>
PGP Key ID: 0x138EA177  (67 29 D7 BC E8 18 3E DA  B2 46 B3 D8 14 36 FE B6)

More information about the NANOG mailing list