Well, I don't really think it would help.

First - I don't understand how it would enable sharing the data among processes/machines.

Second - there are 4228250625 possible questions I might ask (IP addresses) and >65000 answers I might get (AS numbers) - so we're back to 'hash of all IP addresses in the world' thingie, and as an ADDITION to data I already have. And not sharable. Not a good one, no.

But thanks anyway :)