Hi Darren and list, First of all I would like to excuse me for sending part of our private mails to the mailing list. I shouldn't have done it, sorry. I've deliberately started a new mail this time.
Surprise: I'm now mostly convinced about BEEP and its use in application protocols. I've reread some of the RFCs and also read a couple of drafts. BEEP over SCTP is the future, BEEP over TCP is just a nuisance while SCTP is being implemented (SCTP supports multiple streams per connection) This means that I'll do my best to implement RFC3195 in our syslog implementation (syslog-ng). I have a single question left however (to address the legacy problems), is the working group interested in an update of rfc3164 or syslog/sign with the possibility of TCP transport, and the RFC3339 timestamp? My reasoning is: * solve possible TCP interop problems with current implementations * to provide a way to send reliable messages for small edge devices which probably will never implement SCTP nor BEEP (DSL routers for instance) This means that implementors would have three options: * do nothing, log messages are sent via UDP * do the trivial update to TCP * do the full blown syslog/reliable implementation The second option is still much easier to achieve for most vendors, and we would still have some reliability. Current syslog-over-TCP implementations will have something to conform to while RFC3195 implementation is done. What do you think? -- Bazsi PGP info: KeyID 9AF8D0A9 Fingerprint CD27 CFB0 802C 0944 9CFD 804E C82C 8EB1
