home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cis.ohio-state.edu!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!ames!lll-winken!framsparc.ocf.llnl.gov!booloo
- From: booloo@framsparc.ocf.llnl.gov (Mark Boolootian)
- Newsgroups: comp.protocols.tcp-ip
- Subject: Why UDP was slower than TCP using ttcp
- Message-ID: <134530@lll-winken.LLNL.GOV>
- Date: 26 Aug 92 22:06:29 GMT
- Sender: usenet@lll-winken.LLNL.GOV
- Organization: Lawrence Livermore National Laboratory
- Lines: 23
- Nntp-Posting-Host: framsparc.ocf.llnl.gov
-
- Someone asked me to summarize the answer to my question: Why, when running
- over FDDI, was UDP slower than TCP for a Cray X/MP sending to a Sun 4/680 (
- using ttcp)?
-
- I left out one piece of information which appears to be the answer: I did
- *not* fire up a receiver. I simply sent from the Cray to another unsuspecting
- machine. After posting and getting a few responses (each of which talked
- about over-running the receiver), I reran my tests with a receiver in place
- and, sure enough, the data rates were as one would expect (I measured the
- Cray sending around 64 mbits/second). My guess is that sending to a machine
- with no receiver was causing the receiving machine to generate ICMP
- port-unreachable (or some such) error messages, and these were slowing down
- the transfer. Admittedly, this seems a little thin in terms of an
- explanation (at least it does to me), but it sure sounded convincing to
- those around me asking me to explain what the heck was going on...
-
- thanks,
- mb
- --
- Mark Boolootian booloo@llnl.gov +1 510 423 1948
-
- "I don't know that atheists should be considered citizens, nor should they be
- considered patriots. This is one nation under God." - George Bush 8/27/88
-