<No, capacity is limited by the total latency, isn't it? If we arranged
No, it's latentcy * data rate.
For example 10MBS = 10,000,000 bits per second, where if it were
if it were 100,000bits second and the path length were 1 second in both
cases you can see the difference.
Also for the case presented the aggregate storage is path delay times
SLOWEST data rate for path. If there is buffereing there is a more complex
calculation but it will never be MAX data rate.
Of course the possible delay length is far to variable do the dynamic
routing.
Allison
Hi,
I just picked up a couple of Compaq Portable II computers that I thought
I'd pass along to anyone that's interested. These are the second model and
have an 8 MHz 286 CPU with one floppy drive and one (factory) 20 Mb hard
drive. They're in the original travel cases and both the computers and
cases are in nearly perfect conditon. The original power cord is even still
in it's storage compartment. Works PERFECTLY!!! No screen burns, bad keys,
etc. I'll take $50 each plus shipping from 32765. No manuals. Contact me
directly if interested.
Joe
Sorry for the post about a machine too young for the list but it is the
first of its kind in my collection. The chips seem to be from 1994 and the
RiscOS (v3) splash screen also says 1994.
This machine came to me as a box only having been an insurance write off
because it jumped downstairs but there only seemms to superficial damage.
It boots OK but the mouse which appears to be necessary in order to use the
OS is missing and appears to have a proprietary connector. Can anyone tell
me either if the mice are available? Alternatively is it possible either to
persuade it to work with a serial mouse or to wire in a PS/2 socket in
place of the original one so that I can use a PS/2 mouse?
it seems to have 4M of RAM as 1 72 pin SIMM and 1 other memory board which
looks like a miniature DIMM. It also has a HDD with 59 Mb used and 159Mb
free so it must be about 220Mb altogether, this make windows look a bit
bloated as there appear to be some apps there as well - it also seems to
boot very fast compared with modern machines.
Does anyone know what processor it has and how fast it is?
Thanks
Pete
--- John Foust <jfoust(a)threedee.com> wrote:
>
> ... the other night... I remembered a wire delay line memory
>
> Sure, it would be abusive of the Net's resources. You'd need to contend
> with the possibility of dropped packets...
>
> Imagine a chain of machines or routers or whatever that would simply
> pass a special kind of packet to another machine, echoing and mirroring
> packets back to my machine.
>
> ...wouldn't we create a long delay line
> with large data capacity? Obviously the speed of access is nothing
> like a hard drive or RAM, but it would be a neat hack, no?
This very concept came up for lengthy discussion when I was in college at
Ohio State. Then, we envisioned using UUCP to store and forward two weeks
of data, based on the known propagation rates from node to node (since some
would forward messages only after 23:00 to save on toll charges).
We never implemented it, but we did have a theoretical model that probably
would have worked. As you say, access times were hell, but you could store
lots of data if you absorbed hard drives all over the planet.
-ethan
_________________________________________________________
Do You Yahoo!?
Free instant messaging and more at http://messenger.yahoo.com
Actually, I do believe they both supported the length of ECC suitable for
correction, but only the 2010 could interact properly with the 1014 to
perform the correction. Why the 1014 was required in place of host-driver
code, I can't say. One could set the ecc length to either 32 or 56 bits,
and the latter was required for correction. I didn't figure this out until
the details on the PC-based WDFMT program were published, indicating that
the 1010 would support either length, though it was common knowledge that
only 2010 was actually error-correction-capable. It was never clear to me
how the 1014 was going to help with the correction. Apparently it became no
less burdensome to effect correction in the host interface code,
particularly when the scotched the whole notion of bothering with
correction. I don't recall any controller actually going ahead and doing it
as part of the drivers. Somebody did publish an error scrubbing app-note
though.
Dick
-----Original Message-----
From: Eric Smith <eric(a)brouhaha.com>
To: Discussion re-collecting of classic computers
<classiccmp(a)u.washington.edu>
Date: Monday, May 10, 1999 9:41 PM
Subject: Re: WD Chip Info
>> WD2010B - This is the HDC chip, similar to the 1010. Does
>I wrote:
>> Drop-in replacement for the 1010, with ECC. Also can support >1024
>> cylinders, helpful with the Maxtor XT1140 and Maxtor XT2190 disk drives
>
>Minor clarification: the chip is a drop in replacement, but if you do
that,
>I think the drive needs to be reformatted, since the 1010 only uses a CRC
>rather then the longer ECC code.
>> This very concept came up for lengthy discussion when I was in college at
>> Ohio State.
I think it's one of those things that crops up now and then. We had a
similar one here at work whereby we were going to expand our token ring
network to create a fatter 'token doughnut' system. Implementation,
sadly, was never started, but the theory was there...
(it's amazing what you can achieve during boring moments at work!!)
cheers
Jules
On Thursday, May 13, 1999 10:16 AM, John Foust [SMTP:jfoust@threedee.com] wrote:
>
> To put myself to sleep the other night, I was thinking about the recent
> discussions about core memory, and I remembered a wire delay line memory
> I'd once disassembled as a teen, and thought about the old mercury delay
> line memories, and then moved it to the Internet.
>
> Sure, it would be abusive of the Net's resources. You'd need to contend
> with the possibility of dropped packets, which might invalidate the entire
> experience. But like the "unused" CPU cycles on your PC, there is a
> great deal of unused bandwidth. Not everyone's pipe is full, and these
> pipes are a form of transient memory.
>
> Imagine a chain of machines or routers or whatever that would simply
> pass a special kind of packet to another machine, echoing and mirroring
> packets back to my machine. By taking advantage of the delays in
> transmitting packets around the world, across fiber lines, under
> the sea, up to satellites, etc. wouldn't we create a long delay line
> with large data capacity? Obviously the speed of access is nothing
> like a hard drive or RAM, but it would be a neat hack, no?
>
> - John
>
Clever...
Steve Robertson - <steverob(a)hotoffice.com>
To put myself to sleep the other night, I was thinking about the recent
discussions about core memory, and I remembered a wire delay line memory
I'd once disassembled as a teen, and thought about the old mercury delay
line memories, and then moved it to the Internet.
Sure, it would be abusive of the Net's resources. You'd need to contend
with the possibility of dropped packets, which might invalidate the entire
experience. But like the "unused" CPU cycles on your PC, there is a
great deal of unused bandwidth. Not everyone's pipe is full, and these
pipes are a form of transient memory.
Imagine a chain of machines or routers or whatever that would simply
pass a special kind of packet to another machine, echoing and mirroring
packets back to my machine. By taking advantage of the delays in
transmitting packets around the world, across fiber lines, under
the sea, up to satellites, etc. wouldn't we create a long delay line
with large data capacity? Obviously the speed of access is nothing
like a hard drive or RAM, but it would be a neat hack, no?
- John
--- Christian Fandt <cfandt(a)netsync.net> wrote:
> Upon the date 09:55 AM 5/11/99 -0700, Marvin said something like:
> >
> >I haven't heard anything about people going to Dayton except for Ethan...
>
> Hi Ethan, Marvin and others who my be there:
> I've suddenly been able to shift my skeds to come to Dayton too...
Cool.
> ...So, Friday I'll be lurking in the fleamarket to catch those DEC docs
> and other needed bits...
These days, there's not much DEC stuff. When I started going in 1982,
you could find things on maybe as much as 5% of the tables. Now, it's
much less than 1%.
In any case, I'll be at work in Columbus on Friday. I'm only going on
Saturday.
Later,
-ethan
_________________________________________________________
Do You Yahoo!?
Free instant messaging and more at http://messenger.yahoo.com
