> > I have been told, first hand by a local Bay Networks CE, that tests in
> > their lab showed a Wellfleet BLN could manage 11 route flaps per second
> > with 80,000 route entries (running BGP) while at the same time keeping
> > the available links saturated.  I suspect this was with only two
> > Ethernet ports active on each end, but still it makes Cisco 7500's look
> > damn silly.  They say the BLN can do this because it has a true
> > multi-processor architecture, and that packet pushing is done separately
> > from BGP processing.
> > 
> > I've yet to see this work first-hand, mind you.
> 
> Does anyone have a pointer to more "concrete" information WRT the
> cisco problem?  I am going to the cisco user conference in Dallas
> next week, and would like to bring this up.  Say, are any other NetBSD'ers
> going to be in Dallas next week?  If so, we should get together for
> beer/pizza/beer...
> 
	Route flaps can be reduced by having lower priority routes to the
	Null0 destination so some route flap dampening code doesn't really
	ruin your day... This is getting to a technical level that probably
	is beyond the scope of the list.

	The traditional vendors all have their problems, non-traditional
	vendors are providing some higher performance solutions since they
	don't have the burden of legacy customers.

	The commercial Internet should become better congestion wise in
	June or July once the private interconnects between major backbone
	providers come on line. The reduction in overall congestion should
	help alot of things. Right now everyone is funneling through one
	or two choke points, having alot more choke points should help things
	out! B^).

		-Rob