On 8/1/02, Thomas J. Hruska wrote: >Attached is the first Internet Draft of the Unified RFC >Protocol. IETF 54 is long over and the mailing lists need >something else to talk about...so here is something to >discuss. > >The IETF Secretariat should note that Shining Light >Productions has implemented the Unified RFC Protocol into >the ProtoNova Web Server providing standardized remote >module support for the server. The IETF Secretariat >should also note that the Unified RFC Protocol is not >necessarily a lock-step protocol (like SMTP, POP3, and >other protocols), although the document is written in such >a fashion. > >The ProtoNova Web Server is available as a Shining Light >Productions commercial product and people are already >using the server given that it has only been 24 hours >since its release. As such, the ProtoNova Web Server >demonstrates that the Unified RFC Protocol is a viable >communication protocol for generic data transfer and data >packaging purposes as well as a potential replacement >protocol for ailing, ancient, and insecure protocols. > >While the document I am submitting is probably not 100% >perfect, it should be close enough for accepting it as a >first draft. > There is indeed a wealth of simplification that could be accomplished with many of the common internet application layer protocols. However this is not the time to fix those problems. No true simplification will result. Every new client would have to be compatible with old and new servers. Every new server would have to be compatible with old and new clients. No matter how elegant the new protocol was defined, implementations would still be additional logic, new cases for network monitors to learn, additional test cases and additional points of failure. The appropriate time to consider consolidation of application layer protocols might be when client and server changes would be required anyway. When the Remote Direct Data Placement (RDDP) work is completed it might be worth considering a consolidation for the RDDP-optimized versions of the applications this proposal cites. Because of the indefinite transition problems, a mere cleanup will only add logic, not subtract it. Combining a consolidation with the work to standardize application specific use of RDDP capabilities would be a different matter.