raph@kiwi.CS.Berkeley.EDU (Raph Levien) writes: : 1. Encrypt the ping, so it looks like so much PGP traffic. A good idea, if a c'punk remailer cannot decrypt a message it should be considerd "down". : 2. Pad it with a random amount of junk (but not _too_ much :-), so : traffic analysis based on size will fail. That would be the start of RemailerNet. :) : 3. Chain it through other remailers. A good approach might be to : choose two random remailers out of the "top five," and sandwich rebma : between them. The drawback is that it penalizes rebma for their : latency and failure rate, but this might be acceptable. One question about your "suggested path", it looks like you thake the tree fastes remailers [if using kiwi.cs.berkeley.edu]. If the two remailers from Hal Finney are consecutive [ie. 1 and 2 fastes] a suggestion to mail from alumi to portal would IMHO be wrong, because for security reasons both remailers should be regarded the same, since they are operated by the same person. [same goes ofcourse for the usura@hacktic.nl and the vox.hacktic.nl remailers wich are operated by me] -- Exit! Stage Left. Alex de Joode <usura@vox.hacktic.nl>