On Sat, Jan 3, 2009 at 4:45 PM, Tony Li <[email protected]> wrote: > |The well-grounded trend lines I posted suggest that the BGP table size > |at the edge of what we can build cost-effectively may be compounding > |at a rate of 40% per year while the actual size of the BGP table may > |only be compounding at around 25% per year. If true, one consequence > |should be that the minimum cost of an edge device capable of > |participating in the BGP backbone has fallen significantly. This > |appears to be the case; a $1400 Cisco 2811 is now capable of > |interacting with the full BGP table while 15 years ago you had to > |spend $60k on a Cisco 7000 to seriously think about a full BGP table. > > More than something of a red herring, given that they are violently > different generations of hardware, with very different levels of hardware > optimization. The 7000, for example, was built before the networking > industry was on to the ASIC bandwagon. Since we've transitioned from > technology laggards to being near cutting edge, there have been some one > time advances.
The history of the computing industry is one of regularly encountered one-time advances. Moore's Law isn't built on the back of purely evolutionary improvement. That today's router hardware is radically different than that of 15 years ago doesn't refute the sustainability of the growth trend; it demonstrates it. > |So no, the truth of the claim that we're growing the table faster than > |technology can keep up is not at all clear. > > Fair enough. > > So, why do we keep debating this? We've been over this to death. The horse > is dead already. For folks that feel that there's no problem, why not leave > the rest of us in peace to do what we feel we need to do? I'm debating it because I'm tired of watching us bounce off the pinball bumpers. Resolving this issue impacts our work here in three very important ways: 1. It affects whether as a group we reject strategy F, and if we don't, the language with which we present it. 2. Engineer's maxim: if it ain't broke, don't fix it. If failure does not loom on BGP's horizon, there is little justification for moving any disruptive strategy forward from research to engineering. The desirable output from our efforts is *very* different depending on whether or not BGP is failing. 3. RIRs make painful, real-world-hurtful decisions based on the credence they lend the warning of BGP's impending failure. If the end of the world is no longer nigh, we have a moral responsibility to let them know. #3 hits me particularly hard at work where an $800k continuing operations component of a multi-million dollar project can't justify a /24 from ARIN for multihoming. However frightful the monsters under the bed may be, there is insufficient proof of their existence to justify giving your toddler a shotgun. Regards, Bill Herrin -- William D. Herrin ................ [email protected] [email protected] 3005 Crane Dr. ...................... Web: <http://bill.herrin.us/> Falls Church, VA 22042-3004 _______________________________________________ rrg mailing list [email protected] https://www.irtf.org/mailman/listinfo/rrg
