I just bought into ANCR as pure speculation. I figure there is a
reasonable chance some company may buy them out for their patents or
enineering.
Now, for my concerns which I post to the group for any technical
insight. My main problem with Ancor is that they have, until about a
year ago, focused almost entirely on Fibre Channel as a network
connection, neglecting the storage side. This is beginning to change.
Several of us have been preaching to them about exploiting the storage
attachment side of Fibre Channel to leverage investments that will be
made in this arena to sell switches. Lots of people will be buying
Fibre Channel to as a disk attachment. Ancor should position the
company to connect into that infrastructure.
This is in fact Fibre Channel's primary weapon against the
competition. With ATM and Switched Ethernet you have
to add expensive hardware in front of the disks before you
can actually attach them to a network; or you have to add
a server. Not so with Fibre Channel: the drives plug directly
into the network. This part of the charm---you just keep
adding disk after disk without paying a fortune for the hardware
to connect it all together. Don't expect Sun to do this for you:
their server sales are too lucrative. This network-attached
storage approach scales well in both capacity and bandwidth
since you can add a switch when necesssary to get the bandwidth.
Now, what clues one would have to see from companies like
Brocade or Ancor if they were truly positioned to make big
bucks? In my opinion, we should be seeing people writing
system software which can exploiting FC. I am talking
optimized TCP/IP drivers, device drivers, and file systems.
Switch companies do not have to do all the labor, but they
better be working closely with the pros hacking OS kernels.
FC can easily support *sustained* file system transfers
of 100-150 Mb/s (bytes, not bits). I have done this myself.
This will support full-motion video for say digital editing
of motion pictures. I have seen plans for systems on the
scale of 20 Gb/s.
Alternatively, stand in the way of Cisco and HP to duke it
out with high-speed network connections, Ancor is going to
be squashed. Just like the bug which gets caught in the
slipstream of a fast moving car, there is nothing the insect
can do to avoid impact with the windshield. Things would
change in a hurry if a 3COM bought the technology. Then FC
*might* have a chance against GE and ATM for high-bandwidth
uplink and interconnect. This is also pure speculation. |
