> I expect server vendors such as Sun, HP and others to start using
> FC/AL, FC switches to connect to disk arrays for increased
> performance. Old disk arrays with SCSI interface will start using
> adapters to fit into the big picture. Ancor will get a healthy market
> share IMO.
I actually agree with you that server vendors will
upgrade to higher performance. I was making a more
subtle point: It is quite possible to have network
storage without a fileserver, per se. Workstation
vendors will fight this as best they can, but the
marketplace will beat them back. The big winners
when the dust settles will be the network integrators
and, maybe Seagate who makes FC drives in production
quantities.
Let me explain. Let's take Sun, for example. Their
servers usually employ Network File System (NFS)
meaning you typically have a Sun workstation in
the mix somewhere. Actually, a wide variety of vendors
will let you use NFS (SGI, DEC, HP, etc.), but Sun
dominates this world.
"Networked storage" does not require this. All you need
is a very fast switch with the right drivers at striped
RAID. This is very important. Let's say you are a
major movie house making a digital effects movie.
To have the SGI server attached to the super-fast
disks, means you have people waiting in a queue.
Studio production schedules cost money by the hour.
A full blown movie might require 2000 Gb of disk
space. Say, $1K/Gb translates to over a two million
dollars for that seat. Plus guys are lining up
waiting to play back their 20 seconds of action.
Imagine the scenario where you check out blocks of
disk space much like books in a library. Now the
whole movie can be assembled on one or two seats
with thirty animators working like crazy on
production. This requires switching at the speeds
we are talking about. NFS will not work because
of the overhead of the protocol.
Ancor had a major technical lead in FC. I think they
still have. They have made FC switches for four years
or so. However, people copy. Bragging rights are
settled at the marketplace, not in R&D Magazine.
For example, Microsoft never wrote a single piece of
quality, original software in the corporation's
history. So they are chided at in technical
conventions, but no one laughs at the bank
or on Wall Street.
I am not privy to Ancor's current business plan.
I believe their original one does not cut the
mustard. My hope is they have dug in and focused
on the right target which must be within their reach.
Killing off ATM is probably off the radar.
Sorry for the long-winded reply... |
