[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [RAM] Number of DFZ routers

Noel,
 
No,no,no. I am talking completely different things. 
At this point in time I just responded to an email which sounded pretty desperately and I only wanted to express my view that there is no reason to be desperate (unless ALL are in favor of BGP forever ,  which I don't think is the case) - well and all of this one day before going on vacation :-)
 
 
In einer eMail vom 07.07.2007 21:28:05 Westeuropäische Normalzeit schreibt jnc at mercury.lcs.mit.edu:
> From: HeinerHummel at aol.com

    >> The size of the FIB is caused by the way IPv4 addresses are assigned,
    >> and has nothing to do with the algorithm/protocol by which paths are
    >> computed.

    > Sorry, I was too sloppy. 2k, mayby 1k of addresses are even sufficient.

Sure, if we can reassign all IPv4 addresses from scratch, for every network
in the entire world. Lots of luck getting everyone to agree to do that.

And then, of course, you have to keep them perfectly assigned with the
connection topology - i.e. almost no PI addresses at all, for anything
smaller than, say one of the 500 largest organizations in the world. So we're
talking maybe the US gets 100 PI addresses, Europe as a whole 150, Japan 75,
the entire rest of the world gets 75, and that leaves us 50 spare.

Note that means all mid-level ISP's don't get their own address space, they
have to get it from one of their upstream ISP's - and if they change
upstreams, all their customers have to change their addresses.
There is no need to reassign all IPv4 addresses from scratch (after all, that would only be perfect for this current moment in 2007). I wouldn't even mind if the address assignment were even worse.
 
Enjoy your Chicago meeting
 
Heiner
 
_______________________________________________
RAM mailing list
RAM at iab.org
https://www1.ietf.org/mailman/listinfo/ram