[dns-operations] "How Dynamic are IP Addresses?" (SIGCOMM, January 2007)

[Sean Donelan]

On Mon, 3 Sep 2007, bmanning at vacation.karoshi.com wrote:
> 	it was a nifty preso - a couple of points that I brought
> 	up w/ the author:
>
> 	a) how does one get "off" the dynamic pool?
> 	b) granularity of pooling algo? - it does hid small static blocks
> 	c) no IPv6 support - e.g. RA/ND is invisable to this approach

Its a good paper, althought I wish people wouldn't assume media type and 
addressing method go together.

There are several plateaus of address stability with manual (usually 
years or less), lease-time (usually weeks or less), session (usually days 
or less), and transaction (usually hours or less) address assignments.
Any address management protocol (manual, dhcp, ppp, pppoe, nat, etc) could 
use any address lifetime, although each protocol tends to have a typical
duration.

A question is how quickly can you detect if a particular address block 
has either increased or decreased its address stability.  Has an address
block changed from a ppp-dial-up pool to a web server with manual 
addresses?

I also wonder if the difference between addresses is the result of 
red-lining.  Are the ratios different because red-lining has forced most
of the "good mail" servers to move to different address blocks?