RE: Proposal: De-boganising New Address Blocks


[me puts the Devil's advocate suit on]

If I read this correctly, the way I summarize it is:

"Instead of waiting for people that get real delegations in a newly
allocated /8 to have real problems, we delegate parts of this new /8 to
pilot projects before hand, knowing that there will be issues, which
allows us to contact operators that are still bogonizing the new block.
The idea being that the inconvenience caused by a pilot project being
bogonized is of little consequence where the inconvenience of a real
delegation being bogonized is real."

Good idea, no contest. Now, the devil's advocate asks: what makes you
think that operators/ISPs are going to react faster to your pilot stuff
being bogonized than they would to real traffic being bogonized, as if
it's a pilot project it's by definition not urgent and can wait

Although I salute the effort, I am concerned that this will not change
current reactive behavior which is to wait for the shit to hit the fan
to update bogon lists. Might sound sad, but I think the way to
de-bogonize new blocks is to automate the de-bogonization process, as
the idea to de-bogonize ahead of time will not cut it as a priority,

My $0.02, and let me stress that I did not intend to sound negative as
the idea is good; I'm just a little skeptical.


Assuming the pilot program does some form of reachability testing and then
some effort is made to notify those with bad filters (good luck), then at
least this notifies them before it's a real inconvenience for anyone.
They may or may not choose to react, but at least this puts them on notice
that they have a problem that will be real in the near future.

It would be good that the statistics be provided how many locations tested
failed routability of previously bogon space. And after end of the testing
and efforts to contact network admins, the retesting should be done and
again similar statistics provided as well as directly list of ips where
at the end the blocks were still not working.