🚀 go-pugleaf

>In article <1991Aug29.041708.29784@unislc.uucp> martin@unislc.UUCP
(Martin Cryer,D1Z01,5754) writes:
>>I think you are asking too much of a timesharing O/S even with an EISA
>>PC to get sustained 6Mbytes/s. If you only want it for bursts, you
>>may have better luck.

In article <1991Aug29.134749.5982@ctr.columbia.edu>
bill@connection.prospect.com (Bill Poitras) writes:
>Are you kidding?  I have seen 100Mbytes/s when ftping a file to a PS/2
>running AIX with a ESDI controller and 16 bit ethernet controller. ...

100 megaBYTES per second over a ten megaBITS per second link is quite
remarkable.  Even Van Jacobson, long acknowledged the TCP speed wizard,
has yet to exceed the hardware bandwidth limits. :-)

Seriously: full Ethernet wire speed is possible; Van has been doing it
for some time.  It requires tuning the software, and in some cases
discarding broken hardware (such as the Intel 82586 Ethernet controller);
it also tends to put a strain on the bandwidth of the narrower busses
(including many workstation or PC memory busses).  Getting 60%, or
6 MB/s, sustained throughput should not be too hard though.

100 MB (megabytes, not megabits) per second is possible, but requires
the ability to do 25 million 4-byte bus cycles per second, or 50
million 2-byte bus cycles per second; most small-machine busses are
simply not up to this.  Of course, it also requires a `wire' capable of
the same rate.  FDDI is 100 Mb (megaBITS) per second: a mere 12.5 MB/s,
well in reach of current workstation capabilities.

Waiting for 100 Gb/s links :-) ,
-- 
In-Real-Life: Chris Torek, Lawrence Berkeley Lab CSE/EE (+1 415 486 5427)
Berkeley, CA		Domain:	torek@ee.lbl.gov
		new area code as of September 2, 1991: +1 510 486 5427