home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!think.com!ames!agate!ucbvax!FTP.COM!jbvb
- From: jbvb@FTP.COM ("James B. Van Bokkelen")
- Newsgroups: comp.protocols.tcp-ip.ibmpc
- Subject: Re: Fragmented IP packets: any PD implementations?
- Message-ID: <9211171706.AA01113@ftp.com>
- Date: 17 Nov 92 17:06:05 GMT
- Sender: daemon@ucbvax.BERKELEY.EDU
- Reply-To: jbvb@ftp.com
- Distribution: world
- Organization: The Internet
- Lines: 17
-
-
- In the real world, fragments are not terribly important...
-
- True, but the commercial vendors encounter all sorts of oddball situations.
- Once upon a time, when PC/TCP was less than a year old, we had a customer who
- had a custom (few-of-a-kind) router that fragmented IP datagrams larger than
- 510 (what a number to pick) bytes. So, in 1987 we put in reassembly that
- would handle two fragments only. That lasted until 1992, when we went the
- rest of the way, to N fragments. Some long-haul nets have *odd* MTUs, and
- not all routers support enough ICMP to do MTU Discovery...
-
- If you do it, read Dave Clark's RFC - it doesn't really take that much memory
- unless you're getting lots of fragments coupled with significant packet loss.
-
- James B. VanBokkelen 2 High St., North Andover, MA 01845
- FTP Software Inc. voice: (508) 685-4000 fax: (508) 794-4488
-
-