--- Kevin Schoedel <schoedel_at_kw.igs.net> wrote:
> >Those numbers are on a par with old MVIIs with DEQNAs and MV2000s (ca1988).
>
> Hey! What's wrong with DEQNAs and MV2000s?
One problem I know about is that they can't handle as much network saturation
as other equipment of the same vintage. At OSU, c. mid-1980s, a friend of
mine calculated the minimum back-to-back packet time that their Pyramids could
fling packets onto the network at. The idea was to raise network throughput
by reducing inter-packet time since most of the traffic was NFS and telnet,
not heavy, sustained, long packet traffic, IIRC.
When the modifictations to the Pyramid's network driver were implemented,
the lone MicroVAX-II crashed. The DEC interface apparently could handle
passing
a packet to the bus while receiving a new packet from the wire, but when the
next packet was coming in and the first packet wasn't fully off the card, the
DEQNA raised some sort of error which caused a kernel panic. In other words,
there was only room on the card for one unprocessed packet, not two. The
real problem was that when that extra packet came along, the DEQNA should have
ignored it, forcing a later retry, not lock up the Qbus. Software wouldn't
have fixed this problem.
I do not know why DEC obsoleted the DEQNA in favor of the DELQA. All I have
is a DEQNA myself and it's so new (to me), I haven't even plugged it in.
-ethan
=====
Infinet has been sold. The domain is going away in February.
Please send all replies to
erd_at_iname.com
__________________________________________________
Do You Yahoo!?
Bid and sell for free at
http://auctions.yahoo.com
Received on Mon Nov 15 1999 - 09:54:20 GMT