On 11 Jun '08, at 5:28 PM, james woodyatt wrote: I think what Mr. Murphy meant to say was that you're well on your way toward reinventing TURN for TCP <http://tools.ietf.org/html/draft-ietf-behave-turn-tcp>.
Perhaps. If he had said that, it would have been just as inaccurate. I would add that it looks like you might be trying to reinvent both APEX [RFC 3340-3343] and SACRED [RFC 3760]. Good luck with that.
I can see that APEX might seem similar if one weren't familiar with P2P systems. I'm certainly aware of it, having been contributing to the IETF IM discussions ten years ago while Marshall was proposing APEX for use as a standardized IM protocol. And I was using BEEP until a month ago. The differences are pretty fundamental, though — APEX is client/server while Cloudy is P2P. Cloudy uses distributed PKI, but no global names of any kind, not even DNS. APEX uses point-to-point message delivery while Cloudy uses a limited form of epidemic/gossip flooding.
Some protocols that Cloudy actually _is_ a lot like (quite intentionally) are Unmanaged Internet Architecture and SDSI, as well as some of the work done by Steve Dohrmann and Carl Ellison at Intel Labs.
In general, these assertions of what I'm "reinventing" are coming off as rather smug and arrogant. As engineers we all feel ourselves capable of making snap judgments about the worth of anything that another engineer has spent months on; but it's best to recognize that as a fallacy, count to ten, and move on. I have, honest, spent three years researching this area, including about a year of full-time development of various prototypes.
I'm sure there are some RFCs or other protocols I've overlooked, and I welcome pointers, but it would be more polite to say "___ did something like this" or "have you looked at ___?", instead of immediately presuming you know more than I do about my own project.
Anyway. I got my answers about the specific ABS question I asked about, for which thanks; so I think we can close this thread now.
—Jens |