: I have no problem with the feasibility of taking a disk drive and adding an
: embedded general purpose processor to make an integrated system. It should
: be overtly obvious to anyone this is possible and maybe even a winning
: product.
:
: The original proposal from Larry McVoy was to replace the electronics now
: inside a disk drive with an embedded general purpose processor.
[...]
: Now I have been addressing that proposal as not feasible. And it is not.
: Note he is specifically stating he envisions the embedded general purpose
: controller actually replacing the standard disk drive electronics.
First of all, you're picking nits. Second of all, I didn't say
"replace the electronics inside the disk drive", I said "On the disk
drive, the embedded 680x0 is gone, replaced with an embedded Celeron".
That leaves a huge chunk of the electronics intact. Third of all, the
picture I painted is an end point. The drive venders will obviously go
through a progression to get there (notice that Quantum bought Meridian?).
If the market will support a $1K drive + CPU + NIC, then putting the CPU +
NIC down on the drive directly is just a standard cost reduction.
Finally, the whole point is to end up with one sort of cable used
for connecting machines to machines and machines to drives. Yeah,
it's a stretch, but that is the sort of thing that gets my attention.
This all reminds me very much of when I was running around Sun saying that
we ought to be trying to make ethernet run at 100Mbits. At the time,
people like Karen were calmly explaining how there was no standard, how
the industry didn't work that way, how FDDI was better anyway, etc., etc.
Didn't slow me down one bit. There will always be people who stand in the
way of change - they fear it, they have investment in the status quo, etc.
Instead, we started to quietly look into it, Crescendo with their
CDDI showed us that 100Mbit over copper was doable (they were shipping
product way before people started working on 100baseT), and we (Sun)
got pretty jazzed about it. In the early days, it was just me and Andy
B. talking to various companies, not as Sun people, just as engineeers.
It pretty quickly took on a life of its own and I lost interest once it
was clear that it was going to happen.
Same thing here. While today the 100Mbit ethernet looks like a no
brainer, it was not clear at all at the time to the unwashed masses that
it was going to be a success. I think we're at the same stage with the
disk drives. So as long as there is a large installed base of nay sayers
and no real industry support for this, I'll be out there saying (over
and over) that it can be done and should be done. It _WILL_ happen.
Remember that I said that (somebody will research this in 5 years and
I want 'em to find this message in Deja News).
The thing that worries me much more than whether or not it will happen,
is whether or not all the drivers talk to each other with something
better than good old TCP. Something like STP is very obviously needed
and if the SGI STP/Linux is any good at all, I'll push it like crazy
as a standard for the disk drives. Since I'm just one voice in the
wilderness, you all should be thinking about this and pushing it as
well if you decide it's a good plan. Otherwise, what we end up with is
Maxtor's "Accelerated Network Attached Disk Protocol" which doesn't talk
to Quantum's version, etc. That shrinks the market dramatically.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Feb 15 2000 - 21:00:25 EST