home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!elroy.jpl.nasa.gov!usc!sol.ctr.columbia.edu!hamblin.math.byu.edu!hellgate.utah.edu!cc.usu.edu!jrd
- From: jrd@cc.usu.edu (Joe Doupnik)
- Newsgroups: comp.protocols.tcp-ip.ibmpc
- Subject: Re: NDIS (was: TCP/IP and Windows for Workgroups)
- Message-ID: <1992Nov23.214742.61231@cc.usu.edu>
- Date: 23 Nov 92 21:47:41 MDT
- References: <1992Nov21.071344.19293@alisa.com> <TARJEIJ.92Nov23094109@ulrik.uio.no> <1992Nov24.031205.25101@alisa.com>
- Organization: Utah State University
- Lines: 33
-
- In article <1992Nov24.031205.25101@alisa.com>, denny@alisa.com (Bob Denny) writes:
- > In <TARJEIJ.92Nov23094109@ulrik.uio.no>
- > tarjeij@ulrik.uio.no (Tarjei Jensen) writes:
- >
- >> A local wendor claims that ODI is more reliable than either NDIS or packet
- >> drivers. This is especially true under MS Windows. The person I talked to
- >> claimed that he'd never experienced "hanging" with ODI. Supposedly one does
- >> not have to worry about what communication medium one uses when using ODI.
- >> [...etc.]
- >
- > As I said in the abovereferenced article, I have used the he-- out of NDIS
- > with combinations of DECnet, TCP/IP and LAT terminal services... No hangs.
- > It just Works. I suspect that the claim that ODI is "more reliable" is
- > based on the fact that "nobody" but Novell uses it. Has your vendor
- > actually tried to run heavy traffic over an ODI-based Ethernet connection
- > with two or more different protocols at the same time? It is theoretically
- > possible...
- > _______________________________________________________________________________
- > Robert B. Denny voice: (818) 792-9474
- > Alisa Systems, Inc. fax: (818) 792-4068
- > Pasadena, CA (denny@alisa.com, ..uunet!alisa.com!denny)
- --------------------
- Does depend. I've had plenty of trouble with traffic entering a
- Lan Man NDIS stack for AT&T StarGROUP where the sender just pumped packets
- at a merry rate. The stack crashes, but I don't point the finger at the
- NDIS part. On the other hand ODI has performed well here. ODI does make
- provision for queueing requests and if the application hands out the buffers
- the traffic can hit some nice peaks. I do beat up on my own ODI code in an
- application (joint IPX and TCP/IP, same board, hammering away) and it's happy
- too, if that is of any value. All this says is particular implementations of
- this stuff can be good or not quite so good, and the higher layers can't be
- ignored either.
- Joe D.
-