- NANOG - lists.nanog.org

Re: "routing table slots" and the real problem
by Scott Bradner 02 Mar '97

02 Mar '97

Mike says -- The REAL problem is the growing complexity of the ROUTING COMPUTATION, not the size of the resulting forwarding table. even if routers had infinite memory, we would still be crushed by the routing computation if allowed to grow unchecked. -- and the frequency at which the computation must be done Scott

1 0

Re: "routing table slots" and the real problem
by Paul Ferguson 02 Mar '97

02 Mar '97

At 12:09 PM 3/2/97 -0500, Mike O'Dell wrote: > >The REAL problem is the growing complexity of the ROUTING COMPUTATION, >not the size of the resulting forwarding table. even if routers >had infinite memory, we would still be crushed by the routing >computation if allowed to grow unchecked. > I have always subscribed to this fundamental concept; memory to accommodate N number of routes has never been an operational issue. As you have stated, it's the computational resources required to calculate optimal paths when path information changes. There is a linear relationship between the number of paths & routes that must be parsed, and the amount of time it takes to compute install the best one(s). - paul

1 0

Re: The Big Squeeze
by Paul Ferguson 02 Mar '97

02 Mar '97

At 11:33 AM 3/2/97 -0500, Jim Jagielski wrote: >> >> Que sera, sera. Renumbering is a fact of life. >> >> See: RFC1900, RFC2008, RFC2071. >> > >Never said it wasn't a fact a life, just that it's a painful >one... And a disruptive one. Imagine the heartburn if a group with >simply one ClassB was required to totally renumber to another... > I never meant to imply that it wasn't disliked and painful. One also might suggest that address portability is very much dependent on the prefix size and it's ability to aggregated (or not) elsewhere in the global topology. Having said that, a /16 is generally considered portable. - paul

1 0

Re: The Big Squeeze
by Paul Ferguson 02 Mar '97

02 Mar '97

At 08:14 AM 3/2/97 PST, Randy Bush wrote: >User-defined paradigm shifts. I bet that is what the Loony Sociopaths for >Demagoguery (acronym intended) are trying to achieve on a number of these >lists. I suspect it may require more experienced and rational users. One >shifts paradgms with experience and consensus, not volume. > 'Rational' being the operative word here. ;-) - paul

1 0

Re: The Big Squeeze
by Paul Ferguson 02 Mar '97

02 Mar '97

At 07:51 AM 3/2/97 PST, Randy Bush wrote: > >If the ops are heading toward the above-described convention, would it not >be cool if the router vendors made it the default? > Paradigm shifts can happen anytime. Best to make it user-definable. :-) - paul

2 1

Re: The Big Squeeze
by Paul Ferguson 02 Mar '97

02 Mar '97

At 07:36 AM 3/2/97 PST, Randy Bush wrote: >> I would suggest that the largest percentage of flapping prefixes in the >> global routing system belong to prefixes longer than /19. > >Hence the convention to damp differently for different lengths. See one of >the foils in http://www.psg.com/~randy/970210.nanog/, which suggests that we >over here start following the European lead on this. > Also, the dampening defaults: bgp dampening <halflife-time> <reuse> <suppress> <maximum-suppress-time> <halflife-time> default is 15 minutes <reuse-value> default is 750 <suppress-value> default is 2000 (I thought it was 1000, but docs indicate otherwise) <max-suppress-time> default 4 times halflife-time - paul

2 1

Re: The Big Squeeze
by Paul Ferguson 02 Mar '97

02 Mar '97

At 07:36 AM 3/2/97 PST, Randy Bush wrote: >> I would suggest that the largest percentage of flapping prefixes in the >> global routing system belong to prefixes longer than /19. > >Hence the convention to damp differently for different lengths. See one of >the foils in http://www.psg.com/~randy/970210.nanog/, which suggests that we >over here start following the European lead on this. > This appears reasonable to me. - paul >randy > >

1 0

Re: The Big Squeeze
by Paul Ferguson 02 Mar '97

02 Mar '97

At 11:58 PM 3/1/97 -0500, Craig Nordin wrote: > >Isn't there a way, if the InterNIC and the larger backbone operators >cooperated, that organizations having smaller armounts of address space >would not be filtered out? > >Or is it technically impossible? > This isn't so much an issue of warm-fuzzy technical fairness, more than it is one of provider (interior) network stability, and when it comes to the latter, the providers who are filtering on prefix length are doing so because they feel that it is in their best interest. I would suggest that the largest percentage of flapping prefixes in the global routing system belong to prefixes longer than /19. This is not to say that they could be economical incentivized to accept routes for arbitrarily long prefixes. US$.02, - paul

2 1

Re: The Big Squeeze
by Sean Donelan 02 Mar '97

02 Mar '97

X-News: sdg.dra.com dra.mail.nanog:7843 >The address allocation scheme is geared towards trying to promote >utilization of IP space, thus the sorta "take just what you >need" methodology. > >The filters that you talk of seem to me to be crude >proxies for controlling routing space on a particular >providers network, this seems to me to be a reasonable >thing (i.e. they have to make their network work). Except the current allocation practices seem at odds with the goal of minimizing route table growth. Why is it better to allocate several non-agregatable blocks that are 'just' the right size rather than one aggregatable block the next size larger? So which do providers really want to minimize, the number of route entries or the size of individual route entries? -- Sean Donelan, Data Research Associates, Inc, St. Louis, MO Affiliation given for identification not representation

3 2

oops
by Sarah Noelle Pratt Ferguson 02 Mar '97

02 Mar '97

worldnet.att.net is the correct domain. -s-

1 0