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

Conference netcad::hub_mgnt

Title:DEChub/HUBwatch/PROBEwatch CONFERENCE
Notice:Firmware -2, Doc -3, Power -4, HW kits -5, firm load -6&7
Moderator:NETCAD::COLELLADT
Created:Wed Nov 13 1991
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:4455
Total number of notes:16761

1818.0. "Is DECconcentrator perf on the WEB ??" by ANNECY::MAIGRET_Y () Fri Dec 23 1994 03:22

    Hello,
    
    I had a look in the dechub WEB, to get switches performances, and it is
    very interesting. I am looking now for the same performances but for
    the DECconcentrator 900 MX, and specially for the latency values. 
    
    Are there the same ?
    
    Is it available on the WEB ?
    
    Thanks for your reply.
    
    Regards,
    Yves.
    
    
T.RTitleUserPersonal
Name
DateLines
1818.1NETCAD::SLAWRENCETue Dec 27 1994 09:554
    
    The DECconcentrator isn't a switch; there are no comparable numbers -
    the latency is effectively zero.
    
1818.2what about Bridges, Term Servers ?ANNECY::MAIGRET_YWed Dec 28 1994 03:5012
    Hi,
    
    Thanks for your reply, but I need a clarification.
    
    Can I considere that the "latency values" are known only for the switch
    family, and are zero for all others DECHUB equipements like bridges, 
    Terminal servers, Concentrators etc.
    
    Regards,
    Yves
    
    
1818.3NETCAD::SLAWRENCEWed Dec 28 1994 09:5421
    
    You need to define what you mean by 'latency' for each type of device;
    for a switch, it is typically used to describe the time between the end
    of the receipt of a frame on one interface until the start of
    transmission on the outgoing interface, or the amount of time added by
    the switch over what the packet would take if there were just one
    segment without a switch.  Bridges are switches - no practical
    difference.
    
    Concentrators and Repeaters essentially operate at the bit level (yes,
    this is a slight oversimplification) rather than the frame level, so a
    latency measurement would be in the time delay for a bit (or a symbol
    on FDDI) and is so small that you can treat it as 0 for all purposes
    except figuring cable lengths on your network (the repeater hop rules
    are a practical expression of this latency, and on FDDI it's built into
    the maximum station limit for the ring).
    
    Terminal servers might have many different usefull latency measurements
    depending on how they are being used - questions about them are best
    posted in the TOOK::TERMINAL_SERVERS conference.
    
1818.4Latency *applies* to bridges & switches, not to concentrsatorsNETCAD::BATTERSBYWed Dec 28 1994 10:0210
    The bridges are the switches with their names changed to use the
    latest in marketing/selling ploys. The competition started calling
    their bridges "switches", (of course some of them are real switches
    which don't store-and-forward frames). Our switches do "real"
    bridging (store-and-forward), at wire speed, and have very very 
    competitive latency values. Wiring concentrators do not store and
    then forward frames, thus "latency values" do not apply to a
    network component such as a concentrator, or a terminal server.
    
    Bob
1818.5Thanks for your anwersANNECY::MAIGRET_YWed Dec 28 1994 10:420