- Test - lists.nanog.org

20402 routing entries
by yakov＠watson.ibm.com 15 Apr '94

15 Apr '94

Ref: Your note of Fri, 15 Apr 1994 16:39:45 -0400 Marty, >We will have to incent (carrot) people to do the right thing... Would you be so kind as to share with the rest of us what that "right thing" is. >you have no plan... Would you be so kind as to tell us whether you have a plan ? Comrades Yakov&Peter

1 0

Re: 20402 routing entries
by hoodr＠netcom.com 15 Apr '94

15 Apr '94

>Robert Hood writes: >> I tried. They told me that the forms and policy for removing more >> specific routes were not in place yet, and that there would be an announcement >> about it in a week or so. (I tried to remove 150 routes). I've got >> a few hundred more to do next week. >> >Why dont you just stop announcing them ? I did, but BGP didn't want to let them go (the other end never recieved a network unreachable event). I tried finding a way to remove the routes without reseting the BGP link, but couldn't, so I just reset the links. I hate doing that. Can we change the subject to 20252 routes now? :) BTW: I just sent another set of CIDR NACRs to Merit. I hope to remove another ~350 routes by the end of next week. -- Robert Hood hoodr(a)netcom.com 1-408-983-1510 System/Network Administrator for NETCOM On-line Communications Services, Inc.

1 0

Re: 20402 routing entries
by hoodr＠netcom.com 15 Apr '94

15 Apr '94

>According to NSF/ANSNET data, 23 aggregates ALONG with its covered 898 specific >routes are co-exist in the routing table now. If the ASs withdraws all these more >specific routes, the routing table could be reduced by 898. Given the urgent >need of reduce the routing table size, would these ASs withdraw those specific >routes as soon as possible? I tried. They told me that the forms and policy for removing more specific routes were not in place yet, and that there would be an announcement about it in a week or so. (I tried to remove 150 routes). I've got a few hundred more to do next week. -- Robert Hood hoodr(a)netcom.com 1-408-983-1510 System/Network Administrator for NETCOM On-line Communications Services, Inc.

5 4

None
by Stephen Wolff 15 Apr '94

15 Apr '94

If NSF were to waive the NSFNET Backbone Service AUP **for the sole purpose of assisting providers in effective CIDRization/aggregation to avoid Internet collapse**, would that be of any help? I have alerted the NSF Counsel to this possibility, and they are of course concerned that everyone will cancel their current commercial contracts, switch their traffic to NSFNET, and hurl all the private providers into bankruptcy. I should welcome your comments; if you support the idea (the waiver, not hurtling into ruin), I may also enlist your help convincing Counsel and the IG. -s

7 8

Re: 20402 routing entries
by sob＠hsdndev.harvard.edu 15 Apr '94

15 Apr '94

Marty, You don't seem to like the state all that much. What is your suggestion to get the bridge long enough? (or the river narrow enough) Scott

1 0

Re: 20402 routing entries
by Vince Fuller 15 Apr '94

15 Apr '94

it would have been a reasonable approach had cisco and/or other vendors had a standard release support BGP4 without bugs 12 months ago and had you given 12 months notice C'mon Marty, wake up and smell the kimchi. The writing has been on the wall since RFC 1338 (Supernetting) was published almost two years ago, since shortly after that is when the CIDR addressing plan first started being implemented. cisco and the other router vendors have gone out of their way to support the CIDR effort, having integrated CIDR support into the earliest possible release of their code. CIDR, default, or die - it may not be the law, but it is reality. --Vince

2 1

Re: 20402 routing entries
by henryc＠oar.net 15 Apr '94

15 Apr '94

>In the best of circumstances CIDR reduces the growth of routing >information to roughly log(#sites). So, yes, in theory over time >it could go "to 10 again", but that would require the number of >sites that may be well beyond IPv4 (or even any IPng) addressing >capabilities. Your first 5 words say alot. This assumes alot, including customers that don't switch providers and inject more specific routes (this is happening already). You're also assuming that we have a mechanism for aggregating routes for folks who choose not to (or can't, for some reason). CIDR only buys us a *little* time. Pigs still can't fly :-). henry

2 1

Re: 20402 routing entries
by Peter Lothberg 15 Apr '94

15 Apr '94

> Yes, we have again started trying to get folks to move to BGP4 - heck we > are still trying to get some folks to move to BGP. The results are slow > in coming. > --Elise I knew this sounds hard, but; Offer those who don't talk BGP4, to be disconnected by 15-MAY or so if they don't change. -peter

4 3

20402 routing entries
by yakov＠watson.ibm.com 15 Apr '94

15 Apr '94

Ref: Your note of Thu, 14 Apr 94 19:39:42 EDT Elise, >Yes, we have again started trying to get folks to move to BGP4 -- heck >we are still trying to get some folks to move to BGP. Trying to get folks to move to BGP4 is a good thing to do and it should be encouraged, but given the time constraints this is unlikely to be a feasible solution to the problem in the near future. >The results are slow in coming... It seems that some folks PERSISENTLY missed the point. The clock is ticking. The only way to slow down the clock is by reducing the volume of routing information (via CIDR). Nothing else is going to stop or slow down the clock. The clock cares less whether "the results are slow in coming" or why they are "slow in coming". And, by the way, when the clock handle hits the mark there is going to be a massive crash. Forgive me my bluntness, but given where we are today I just don't see any feasible alternatives to a medium-large scale proxy aggregation within a next month or so. So, given the finite amount of human resources we have, let me suggest that the CIDR efforts should be strongly focused on the following two FEASIBLE choices: (a) make sites who have BGP-4 TODAY to withdraw individual components asap (b) make sites who don't have BGP-4 TODAY to arrange proxy aggregation by their providers (who assumed to have BGP-4) asap (or the providers will arrange it for them). Yakov.

4 3

20402 routing entries
by yakov＠watson.ibm.com 15 Apr '94

15 Apr '94

Ref: Your note of Fri, 15 Apr 1994 10:21:39 -0400 Henry, >But even in the best of circumstances, all CIDR does is slow the >RPM of the clock from 10 to 2, an over time it goes to 10 again. In the best of circumstances CIDR reduces the growth of routing information to roughly log(#sites). So, yes, in theory over time it could go "to 10 again", but that would require the number of sites that may be well beyond IPv4 (or even any IPng) addressing capabilities. It may also be well beyond any practical requirements for a max size of an internet. Yakov.

1 0