home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.protocols.tcp-ip.ibmpc
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!sgiblab!swrinde!news.dell.com!natinst.com!cs.utexas.edu!sun-barr!ames!eos!ssmith
- From: ssmith@eos.arc.nasa.gov (Stephen Smith)
- Subject: Re: Fragmented IP packets: any PD implementations?
- Message-ID: <1992Nov11.215152.20095@eos.arc.nasa.gov>
- Keywords: IP fragments TCP/IP
- Organization: NASA Ames Research Center
- References: <1992Nov8.203621.6902@mnemosyne.cs.du.edu> <BxGtps.BHM@watserv1.uwaterloo.ca>
- Distribution: usa
- Date: Wed, 11 Nov 1992 21:51:52 GMT
- Lines: 48
-
- erick@demorgan.uwaterloo.ca (Erick Engelke) writes:
-
- >> dpayne@nyx.cs.du.edu (Dirk Payne) writes:
- >>I have reviewed WATTCP, PCIP, and both CUTCP and NCSA's versions of telnet
- >>but none of these implemetations handle IP fragments. In order to be a "real"
- >>TCP/IP package (connected to the "real" world) fragments cannot be ejected
- >>out of hand.
-
- >The question I have is why you are needing frags. There are only a
- >few situations which should generate fragments in a user's PC, large
- >UDP packets and incorrectly configured TCP systems.
- The "few situations" may be important to the user(s), despite what I
- think. Nor is it fair to assume any system which generates them is
- "incorrectly configured" or uses large UDP packets (e.g., slip over a
- noisy line).
-
- I don't think "needing frags" is the point, its a matter of
- co-existence w/ systems with frags-- regardless of how they are
- generated.
-
- [... dispell stuff deleted]
-
- >I might be convinced to add frag support, but would appreciate
- >opinion if it is a good idea, in particular, which applications
- >could use it. NFS is not a great answer since it would be useful
- How about ftp, nntp, smtp, telnet --- i.e. all the WKS. The apps will
- follow once the foundation is built. Several of these exist in NOS.
- >even with 1K packet sizes, the problem there is the lack of an
- >implementation. I know the Host Requirements RFC tells us what
- >to do, but I would prefer to see an application which needs it.
-
- From my end, the problem is various PC tcp/ip implementations that
- don't supporting fragment reassembly trying to co-exist with systems
- that do support reassembly. No one person has control over every
- system, and tcp/ip is the primary protocol stack and pre-dates PC's
- altogether.
-
- My only question are commercial apps and Phil Karn's NOS to be the only
- answer? Given the clammoring for SLIP, the answer is getting more
- important every day.
-
-
- -Steve
- --
- ==================================================================
- Stephen Smith MS 262-6
- (415) 604-1381 NASA Ames Research Center
- ssmith@eos.arc.nasa.gov Moffett Field, Ca. 94035
-