Hi ris-int, since nanog's reaction to my post explaining our AS-set stuffing techniques generated fewer flames than expected, I was thinking of announcing AS-set stuffing experiments over IPv4 today or tomorrow. I would start with an experiment to determine how far large AS-sets penetrate into the Internet. Idea: announce progressively longer AS-sets containing only AS 12654 (so AS-paths like 12654 {n * 12654}) and see if the large AS-sets survive all the way to the RRCs or not. This would: a) provide results we need b) eliminate some opposition by showing that large AS-sets are safe c) not irritate people because we're only using our own ASn The values of n would be 5, 10, 20, 50, 75, 100. This is well below the safety limit we have seen with Ciscos (254). Once that is done, I would announce other experiments using other people's ASes, saying that people can opt-out if they don't want their AS to appear. Thoughts? Cheers, Lorenzo -- --------------------------------------------------------- Lorenzo Colitti Ph.D student Computer Networks research group Roma Tre University colitti@dia.uniroma3.it +39-0655173215 ---------------------------------------------------------