|  |     
    The rate limiting feature (applies to all 900 series switches) works
    like this:
    
    You specify a maximum forwarding rate for rate-limited traffic, in 100s
    of packets per second.
    
    You then configure a filter for each multicast address that you want
    rate limiting to apply to (where the all-Fs broadcast address is just
    another multicast address).  In HUBwatch there is a little icon of a
    clock with an arrow through it that designates whether or not a filter
    is rate limited.
    
    Each rate limited address or protocol is throttled to the maximum
    forwarding rate you configured (the same maximum rate applies to each
    filter individually, but you can configure only one rate for the entire
    box).
    
    The limit is applied as a rate, so you don't just turn your broadcast
    traffic into a square wave; I don't know the details of the algorithm,
    but the effect is that a constant stream of packets on one port turns
    into a steady but slower stream of packets on the others.
    
    I configured these on the bridges that we use for our own network here
    in DEChub engineering, and they've worked very well.
    
    Sales note: I think this is an important feature to point out in any
    FDDI/Ethernet sales situation, because of the potential of a problem on
    the FDDI side to completely swamp the Ethernet side (and our switches
    are fast enough to pass the problem through as fast as the Ethernet can
    take it - which is a feature in itself).
    
 |