Eugene wrote:

A couple of questions to resident cluebots:

1) what is the bulk of traffic originating on user side and which does not end within the ISP?

2) how much of that traffic is in clear? Specifically, what is the percentile of SSL sessions, and what is the trend? Do we at all already see things SOAPy, .NETy, or XML-RPCy out there in the wild?

3) browser-web server connect is bidirectional. Which do's and don'ts needs one follow, if one wants to implement an unblockable (well, an ISP not granting web access won't stay an ISP for long) p2p infrastructure on top of that?

(should have paid more attention to the REST thing on FoRK, damn).

Some illuminating comments highly appreciated.

You might want to check ResearchIndex, the NEC Research Institute's Scientific Literature Digital Library. They have a whole treasure trove of relevant PDFs online: an excellent resource in general(if you haven't come across it already.) Good luck! ~Faustine. http://citeseer.nj.nec.com/cs