home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!zazen!news
- From: stevens@vms.macc.wisc.edu (PAul STevens - MACC - 2-9618)
- Newsgroups: comp.os.vms
- Subject: Re: Vax to Ultrix delay over decnet
- Message-ID: <1992Jul23.174522.7714@macc.wisc.edu>
- Date: 23 Jul 92 17:18:41 GMT
- Sender: news@macc.wisc.edu (USENET News System)
- Distribution: usa,local
- Organization: University of Wisconsin Academic Computing Center
- Lines: 49
-
- In article <1992Jul22.135053.3523@arizona.edu>, leonard@telcom.arizona.edu (Aaron Leonard) writes...
-
- >>Telnet's timeout is generally a few tens of milliseconds.
- >
- >Not true. If it were, then every high- or variable-delay network
- >would quickly be destroyed by TELNET. TCP uses a sophisticated
- >adaptive retransmission mechanism, but RFC-1122 recommends
- >beginning with a retransmission timeout of 3 seconds.
-
- You are probably right. My information comes from watching what
- happens on the network when a packet is lost. I see a retransmission
- in only a few tens of milliseconds. In fact, I have seen them occur
- in only a few milliseconds. This may have happened after the nodes
- discovered how close together they were.
-
- >
- >Your TELNET users must have the patience of Job, then. Performance on
- >all network protocols that I am familiar with will go right down the
- >toilet if you have as much as 2% packet loss, let alone 80%.
- >
-
- Not at all! This I have seen with my own eyes. People call me
- to complain that their Novell server is dead and that their
- VAX/VMS nodes are printing adjacency messages and when I arrive with
- a network sniffer I discover that someone has extended their thinwire
- with 75 ohm wire from Radio Shack. But when I use telnet on a PC,
- I cannot tell that anything is wrong. And I have *seen* that each
- packet is repeated many times before being acknowledged. I am pretty
- sure that NCSA telnet is very simple; it has a timeout value in
- CONFIG.TEL that specifies timout in 1/18ths of a second. Mine came
- with a value of one.
- >
- >A properly configured Ethernet will lose virtually no packets (<0.1%)
- >until it's grotesquely overloaded (>>90%). Greater packet loss than
- >that is sign of electrical misconfiguration or hardware malfunction.
- >Find someone responsible for your network, and encourage him to
- >fix it.
- >
-
- I agree with this paragraph 100%. Unfortunately it is I who am
- called upon to fix about 60 of the ethernet segments on the
- campus. But they are always fixable to the standards you suggest.
- In fact, I like to think I can make it much less than 0.1%. I
- don't think overload should cause packet loss except perhaps
- in software queues on the machines connected to the net.
-
- And thanks to everyone for the hints about the DECNET EXEC parameters
- to adjust retransmission delays. I will try them and if they don't
- work you will hear from my boys. ;-)
-