[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference 7.286::fddi

Title:	FDDI - The Next Generation

Moderator:	NETCAD::STEFANI

Created:	Thu Apr 27 1989
Last Modified:	Thu Jun 05 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	2259
Total number of notes:	8590

695.0. "More performance questions" by JUMP4::JOY (Happy at last) Tue Sep 08 1992 14:10

    I am entering this note for a colleague in Asutralia. Can anyone help
    with the answers? You can contact the autor directly by sending mail to
    Andrew McCoy @MEO
    
    Thanks
    Debbie
    
	I have the following question from a customer relating to FDDI 
        performance.  I doubt that you are in a position to answer it 
        yourself, but hope you will know who to forward it to.

	The customer is the Telecom Jindalee project which as you will 
        know represents sizeable future revenue.

             1.	 Performance
             
             Given a lightly loaded network, what is the minimum time for 
             transmission of the shortest permissible FDDI packet between 
             two nodes using DECnet OSI connected via DACs separated by 
             approximately 10m of fibre ?  The time we are looking for is 
             application to application (ie., complete traversal of 
             stack).  For the purpose of this question assume TP class 0 
             (no retransmission) and a basic message transfer service.
             
             2.	 Latency.
             
             Given the conditions from question 1, what is the latency 
             (start of tx at node A to end of rx at node B) for 
             approximately 0.5 MBytes of data presented at node A's OSI 
             stack.
             
	The intended target environment is Alpha/VMS.  As DECnet/OSI for 
        Alpha/VMS does not yet exist, a theoretical answer, or an answer 
        based on measurements of VAX/VMS or DECnet/OSI for Ultrix would 
        be OK.

	Please do not attack the questions.  If there are inaccuracies or 
        misconceptions in them please point them out, and then make and 
        state any necessary reasonable assumptions with the answer.

	Thanks for your help.

cheers
andrew

T.R Title User Personal
Name Date Lines

695.1 KONING::KONING Paul Koning, A-13683 Tue Sep 08 1992 18:58 14

T.R	Title	User	Personal Name	Date	Lines
695.1		KONING::KONING	Paul Koning, A-13683	`Tue Sep 08 1992 18:58`	14
	I can see some problems answering the question... First of all, the only way TP0 could be used is by running CONS, which while theoretically possible on a LAN is not exactly mainstream, and is unlikely to have good performance. Second, the answer is almost entirely a question of how efficient the software is, which for existing software is open to (possibly substantial) improvement, and for not-yet-existing software may also vary by an order of magnitude depending on the skill of the implementer, the adequacy of the underlying operating system design, and to what extent performance was emphasized in the development process. paul

I can see some problems answering the question...

First of all, the only way TP0 could be used is by running CONS, which while
theoretically possible on a LAN is not exactly mainstream, and is unlikely
to have good performance.

Second, the answer is almost entirely a question of how efficient the software
is, which for existing software is open to (possibly substantial) improvement,
and for not-yet-existing software may also vary by an order of magnitude
depending on the skill of the implementer, the adequacy of the underlying
operating system design, and to what extent performance was emphasized in
the development process.

	paul