[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: CUD and FBD
On 14-sep-2005, at 19:54, Erik Nordmark wrote:
This raises the question whether the packets generated by FBD needs
to be concerned about congestion control.
If we have a simple model of the network where the paths are
symmetric and traffic volume is symmetric, then one can definitely
argue that FBD does no need this, because at most it sends one
packet for every received packet. (Only if a packet has been
received in the last t seconds and the ULPs haven't sent anything
back, will FBD send a keepalive.)
I always imagined that FBD keepalives would adhere to a standard
interval, which would be dictated by the requirements of the
application. For instance, if the application wants to rehome within
15 seconds, FBD keepalives would have to be sent every 10 seconds or
so (when required) which leaves 5 seconds to do the shim magic when
ONE FBD keepalive is missed. Or maybe make this every 5 or every 3
seconds and start action after two or three misses.
It gets easier when the application can survive longer outages.
Are there links that can't handle a fairly small packet every 3 - 20
seconds or so?