[FoRK] filtering /48 is going to be necessary

Adam L Beberg beberg at mithral.com
Sun Mar 11 11:06:35 PDT 2012


Or they could just use geographic prefixing and make the problem trivial.

But that idea is far beyond people who make their living on unnecessary 
complexity.

Instead, we'll have a completely broken system that will soon need more 
upgrades, YAY!


Eugen Leitl wrote on 3/11/2012 1:07 AM:
> ----- Forwarded message from William Herrin<bill at herrin.us>  -----
>
> From: William Herrin<bill at herrin.us>
> Date: Sat, 10 Mar 2012 18:11:09 -0500
> To: Sven Olaf Kamphuis<sven at cb3rob.net>
> Cc: "nanog at nanog.org"<nanog at nanog.org>
> Subject: Re: filtering /48 is going to be necessary
>
> On Sat, Mar 10, 2012 at 5:47 PM, Sven Olaf Kamphuis<sven at cb3rob.net>  wrote:
>> just get routers with enough memory...
>>
>> i'm rather for a "specification" of a minimum supported route-size (let's
>> say something along the lines of 64GB in each border router, it's 2012 after
>> all ;) than for putting limits on the prefix sized announced so "old junk"
>> can still stay connected to the internet.
>>
>> let's say, there is 6 billion people in the world.. if they all have 1 route
>> table entry (average ;) i see no technical limitations on anything produced
>> AFTER 2008 actually.
>>
>> stop buying crap without sufficient ram, or just scrap it and get new stuff.
>> (which you're going to have to do to efficiently route ipv6 -anyway- at some
>> point, as your old stuff, simply doesn't even loadbalance trunked ethernet
>> ports properly (layer 3 based) ;)
>
> Sven,
>
> A) 7 billion people in the world, not 6.
>
> B) 7B IPv6 routes won't fit in a 64GB radix tree once, let alone the
> several times they'd need to in order to be useful in a router. For
> that matter, 6B routes won't fit either. (Hint: FIB plus at least one
> RIB for each peer)
>
> C) Big iron is either using massively parallel FIBs (many copies of
> the radix tree) or they're using TCAM instead of DRAM, a specialized
> tristate version of SRAM. In either case, you're talking 10 to 100
> times the cost, ten times the power consumption and ten times the heat
> versus DRAM.
>
> D) No computer presently exists on which the BGP protocol could keep
> up with today's update rate per prefix with 7B prefixes.
>
>
> A router handling 10M routes is achievable today if we're willing to
> go back to $20k as the minimum cost BGP box. That's an order of
> magnitude more than we have now and three orders of magnitude short of
> where we need to be before we can stop sweating the prefix count.
>
> Regards,
> Bill Herrin
>
>
>
>

-- 
Adam L. Beberg
http://www.mithral.com/~beberg/


More information about the FoRK mailing list