| 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 | 
Objet : AVERAGE TRANSMISSION TIME ON ETHERNET
Hi to you all,
    
    THIS IS NOT A FDDI RELATED TOPIC, but I was told that some of you could
    probabely help me out on this one.  I'll appreciate any help on this
    *****VERY***** difficult one.  It has been cross-posted in DNT note
    2920
I have a customer that want to be assured thet ETHERNET will have an 
average transmission delay of 1 millisecond and that in 95% of cases 
the delay won't be greathet than 5 milliseconds.
    
I have the following numbers
			GAT	  VAL	    PB	      STJ	HMR
Number of nodes		35	  42	    53	      74	96
Throughput (Mb/sec)	.85	  1.0	    1.3	      1.8	2.3
I know that the average packet will be 1KB or 8192 bits.
I know the throughput because I know, for example that HMR will be 
transmitting an average of 2924 packets/sec in worst case.
2924 packets/sec * 8192 bits/packets * 96 nodes = 2,300,000,000 bits/sec
Will it be possible to get an average transmission delay of 1 millisec, 
with a max delay of 5 milli sec in 95% of the time.
I NEED THIS INFO FOR TODAY, AUGUST 8 BY 16h00.
Help guys,
Thanks and see ya at NET U
Andre J Courchesne
632-3747
Distribution :
 
� :   GERRY PYLYSHYN @TRC                  ( PYLYSHYN.GERRY AT A1 AT TRCO01 )
 
CC :  GEORGE ZYZANSKI @OTO                 ( ZYZANSKI.GEORGE AT A1 AT OTOO01 )
CC :  ERIC LEDUC @MQO                      ( LEDUC.ERIC )
CC :  YVES DUFRESNE @MQO                   ( DUFRESNE.YVES )
 
Utiliser l'option LLD pour voir le reste de la liste de distribution
    
| T.R | Title | User | Personal Name | Date | Lines | 
|---|---|---|---|---|---|
| 322.1 | KONING::KONING | Eesti vabaks! | Thu Aug 08 1991 10:39 | 11 | |
| The 1 ms average applies to 50% load, and you're well below that. Should be no problem. (Actually, the 1 ms average is for bridges. The actual network access time is only a part of that, perhaps half. On the other hand, on other node types the system overhead might be higher. Were you talking about bridges? FYI, the Ethernet conference (MOUSE::ETHERNET) is the best place for this question. paul | |||||