Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!usc!wuarchive!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!pacbell.com!ucsd!ucbvax!TERMINUS.UMD.EDU!dzoey
From: dzoey@TERMINUS.UMD.EDU
Newsgroups: comp.protocols.tcp-ip.ibmpc
Subject: Re: NCSA + Novell
Message-ID: <9101081356.AA29302@terminus.umd.edu>
Date: 8 Jan 91 13:55:45 GMT
References: <96246@aerospace.AERO.ORG>
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 28

Your question is basically: "Why can I share a 3C501 card among
non-cooperating applications when this behavior hangs other network
cards?"

The reason you can (sort of) share a 3C501 is that there is almost
no state kept on the card.  This makes it very easy for a second application
to take over the card transparently from the first application 
without forcing a lot of state saving to avoid changing the first applications
environment.

The 3C501 can do this because it is a very simple card.  No buffering, no
scatter-gather i/o.

I wrote (sort of) above because while applications can take turns using
the card, I'm not sure two application can really share the card.  That
is, you may be able to load telnet off your novell network, but I don't
think you could ftp a file to the novell drive.  I may be wrong though
as I haven't tried this.  

So, yes, it works for a 3C501, but the 3C501 is so slow and loses so
many packets that you're usually better off getting a more recent generation
card and using one of the many standard card sharing specs.


			Joe Herman
			U. of Md.

dzoey@terminus.umd.edu