Odds are that the 12H array is not pushing the FW SCSI to its limit, I
could be wrong. I think the fiber I/O and the 15k RPM drives would be
able to push the SCSI's the their limit, and we could possibly see an
increase in I/O's per second, as measured by M.P.E. I wish I had a
diagnostic utility that would be able to give me SCSI statistics, like
"bits per Second", average, min, max, and so on, for each FW-SCSI card
on my machine. Something that would give the info to answer the
question, "Is my SCSI card being used to its fullest potential?".
However, this would not be the main reason to buy into the SCSI to Fiber
Channel solution.

We are going to be using the HP3000 for at least one or maybe two more
years, who knows maybe longer. Should we continue buying old
refurbished, out of production storage devices? No, I don't think so. We
should invest in the most up-to-date storage for our older SCSI servers.
So, why "Fiber"? Doing this will allow our older SCSI servers to share
the same storage solution with new fiber channel severs, and when we do
replace the old servers with new servers that support fiber I/O channels
we still retain our storage solution investment. Besides, we already
have a 15 terabyte SAN, and if one more device (Crossroads SA40) will
allow the HP-3000 to tap into it, why not do it?

If any of you want to look at the SA40
http://www.crossroads.com/Products/ServerAttach.asp and give me your
opinion I would appreciate it. I do appreciate all the opinions that
I've already received, positive and negative, thanks!


>>> Yosef Rosenblatt <[log in to unmask]> 06/14/04 08:22PM >>>
One must keep in mind that one's data transfer rate is only a fast as
the
slowest device one has attached to one's SAN. This is why we speak of
bottlenecks and not bottlebottoms. You are not enhancing the SCSI
speeds by
attaching the SCSI HBA to fibre.

My former employer, EMC, did not support the Fibre Bridge solution on
any
platform. The reason was that they saw little performance enhancement
and a
great deal of trouble in getting all the elements to work. This is to
say
nothing of the finger-pointing between server, disk, switch and bridge
support groups, i.e. it is a support/maintenance nightmare. (I
obviously do
not speak for EMC.)

What is the benefit of the Fibre Bridge solution? Does anybody have
real
performance data that they are willing to share with Mr. Anderson? It
would
be a shame for him to go through with this kludge without at least
knowing
what benefit there may be.

On the other hand I may be completely wrong.

Pray for Peace,
Yosef

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *