[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

2061.0. "TTRT significance for DEFPA?" by OTOOA::JPOND () Mon Jun 10 1996 16:46

    Hi there,
    
    I've seen some reference to adjustment of the TTRT parameter in this
    notesfile, and have tried unsuccessfully to obtain the white paper
    from FILES::NET$ARCH:[PAPERS]FDDI_PERF.PS, so I thought I'd pose the
    question here.
    
    A customer with a big FDDI ring (over 20 Sables with DEFPA's running NT
    Server) is asking about the TTRT and its default time of 8ms.  What
    does Digital recommend with respect to adjusting this parameter in 
    different environments?  This customer was told by someone at Frontier,
    who resells Digital FDDI cards, that they always adjust the TTRT up to
    160ms to avoid (unspecified) problems.  The customer wants to know what
    strategy should be pursued in setting this number: should they be high,
    should the number be different for each member of a ring, should some
    be high and some be low, etc.?
    
    In addition, the customer is asking about the negotiation strategy
    for the DEFPA and the TTRT. Does the negotiation of the time go
    both upwards and downwards (toward a negotiated time)?
    
    Those are his questions. If someone can provide some inputs, I
    would appreciate it.
    
    Regards,
    Jim
    
T.RTitleUserPersonal
Name
DateLines
2061.1Have heard to keep TTRT value close to default....NETCAD::BATTERSBYDon't use time/words carelesslyMon Jun 10 1996 18:4013
    Well, I recall that varying the value of TTRT can affect 
    (sometime dramatically), the ultilization and latency of a 
    very busy FDDI ring, and that it has little affect on utilization 
    or latency on a lightly loaded ring. I don't have any personal 
    experience with changing this parameter from its default of 8ms 
    to an order of magnitude in the range of 160ms, so I can't speak 
    to whether this is useful or not. However I have heard many people 
    more knowlegeable than I of FDDI configs, not recommend changing it 
    too far from its default.
    Perhaps others can chime in here with more empirical evidence of
    the impact of changing it by a factor of 20.
    
    Bob
2061.2NETCAD::STEFANIMon Jun 10 1996 18:5723
    >>                   -< TTRT significance for DEFPA? >-
    
    Most of the DEFEA and DEFPA device drivers offer a switch to change the
    MACRequestedTTRT value whose range is 4-165ms and whose default is 8ms.
    Note the word "Requested".
    
    The TTRT is negotiating via a lowest bidder wins algorithm during the
    FDDI claim process.  If one node asks for 8ms and another node asks for
    4ms, the 4ms. node wins.
    
    Raj Jain wrote a paper on why the 8ms. default is a good idea and
    should be left alone.  The reason many of our device drivers allow you
    to change the requested value is because end nodes are inherently
    difficult to manage.  Some FDDI network managers that want to twiddle
    with the TTRT value indiscriminantly set the end node requested values
    to something very high (eg. 160ms).  Now, for all intents and purposes
    the end node FDDI adapters are out of the bidding process.  They can
    then adjust the requested value at something more manageable (eg. bridge
    or concentrator) and know that they'll win the negotiation.
    
    Generally, this value should be left alone.
    
    - Larry
2061.3OTOOA::JPONDTue Jun 11 1996 11:186
    Larry,
    Thanks for the info.  I looked for Raj's paper at 
    FILES::NET$ARCH:[PAPERS]FDDI_PERF.PS, but it seems to be
    missing.  Do you have a current pointer?
    
    Jim
2061.4NETCAD::STEFANITue Jun 11 1996 13:313
    >>missing.  Do you have a current pointer?
    
    Nope.
2061.512368::thomasThe Code WarriorTue Jun 11 1996 13:451
It's there now (FILES::NET$ARCH:[PAPERS]FDDI_PERF.PS).
2061.6OTOOA::JPONDTue Jun 11 1996 19:182
    Thanks,
    Jim