* marcoh@marcoh.net (Marco Hogewoning) [Mon 09 Aug 2010, 16:30 CEST]:
I think 7 days beyond the length of a conference is too short, it usually takes a while to get things properly routed and being able to announce the prefix while you go that process I think is mandatory to be able to debug. I would set this to a month minimum, maybe with a note that after the conference a maximum of 7 days is allowed for cleanup.
Plus, especially for conferences, a relatively significant amount of time (compared to the length of most conferences) is needed beforehand to have align connectivity, update IRRdbs, have filters updated etc. In my experience this can take a few weeks at least. Likewise, 7 days is short given that in the case of temporary ASN assignments the conference depends on its upstreams to update their IRRdb objects before they can be returned to the NCC.
Second concern is cool off or quarantine of previously used space, you probably want some time in between assignments of these resources just in case. For some situations there might even be privacy concerns with this as a block which was recently used for some experiment all of a sudden starts to carry real data from real people visiting a conference.
Given experience with "debogonising" new IANA assignments, this is a valid concern. The RIPE NCC can alleviate this by assigning in an LRU fashion from the block(s) reserved for this purpose. -- Niels. -- "It's amazing what people will do to get their name on the internet, which is odd, because all you really need is a Blogspot account." -- roy edroso, alicublog.blogspot.com