[ppml] Big numbers

Bill Darte billd at cait.wustl.edu
Tue Apr 8 12:21:50 EDT 2003


But what makes you think the paradigm for address allocation and assignment
will be human interactive RIR structures that exist today?
With an 'infinite' number of addresses that couldn't be exhausted for 1M
years at one per second....let's automate the assignment to the devices that
need them.... estimate the total number of new electronic and electrical
devices that want managing... making queries for assignments.

What is the total number of GET requests per second on the Internet today?

Just seems to me that couching the answers to future problems in the
paradigms of today is the way that we get in trouble.

Bill Darte

-----Original Message-----
From: David Conrad
To: Bill Darte
Cc: 'ppml at arin.net '
Sent: 4/8/03 10:42 AM
Subject: Re: [ppml] Big numbers


On Monday, April 7, 2003, at 03:28  PM, Bill Darte wrote:
> Of course, we are not talking about numbering individuals, but 
> potentially
> every electrical and electronic component as well as subsystem
> perhaps....

And I wasn't talking about individual addresses, but /48s...

>  there is no census data for these things, but undoubtedly this
> represents a very large number as well.

To give a slightly different perspective, if the RIRs were, on average, 
to allocate a /48 per second, 24x7x365, the current global unicast /3 
would last a bit over 1 million years.

Which is good, given how long it would likely take the routing system 
to converge... :-)


More information about the ARIN-PPML mailing list