[rrg] Summary of architectural solution space

[Tom Vest]

Coincidentally, 8:1 is just about the same (max) ratio of loans to  
reserves that most chartered/regulated banks are permitted to  
maintain, to keep them from contributing too much to monetary  
inflation...

And on that note, something entirely different for those needing a  
diversion for the weekend:

http://www.eyeconomics.com/backstage/Presentations/Pages/Bankers_for_BGP_v1.2.html

A shortish explanatory memo:

http://www.eyeconomics.com/backstage/NetStagflationPaper.html

and some illustrative links:

http://www.eyeconomics.com/backstage/Isomorphism.html

http://www.eyeconomics.com/backstage/Isomorphism_M.html

http://www.eyeconomics.com/backstage/Isomorphism_I.html

http://www.eyeconomics.com/backstage/Isomorphism_G.html

Strange but true!

Comments, questions, suggestions, criticisms greatly appreciated...

Tom Vest


On Nov 14, 2008, at 3:48 PM, Noel Chiappa wrote:

>> From: Iljitsch van Beijnum <iljitsch at muada.com>
>
>> the current routing table is 8 x the number of ASes and BGP not only
>> fails to contain updates to a limited part of the network, it  
>> actually
>> amplifies them as they circle the globe. Those issues are separate  
>> from
>> an id/loc overload.
>
> I can't say for sure about the first, because I don't know the exact  
> cause(s)
> of that, but as to the second, I think your conclusion is entirely
> unwarranted.
>
> To fix the 'contain updates to a limited part of the network' issue,  
> an
> topologically organized allocation of 'routing-names' (either  
> hierarchical, or
> landmark routing, or _something_ like that, which allows limiting  
> the scope of
> information about a given destination). Yes, it's true that the  
> routing system
> _at the moment_ is incapable of using any such well-organized  
> routing-names -
> but that doesn't obviate the *requirement* for them, if one _does_  
> want to
> limit the scope of updates.
>
> And here we are precisely back to the need for routing-names which  
> aren't
> permanent, _uncorrelated to location_ - because it's evident that  
> the users
> *do* need permanent names for hosts, so there's a direct conflict  
> between what
> users want (permanent names) and what routing needs (names which  
> change when
> location changes), and there is no way to square that circle with a  
> single
> namespace.
>
> 	Noel
> _______________________________________________
> rrg mailing list
> rrg at irtf.org
> https://www.irtf.org/mailman/listinfo/rrg