Mladen Turk wrote: > Not even that. We are talking for more then a year for > a next generation binary http(s) protocol. > > Almost everyone agreed that we need > at least few things: > 1. Encryption > 2. Variable sized messages > 3. Client connection close notification.
Talk about a hijaak :) I'm going to argue; 'no', but let me offer my rational... 1. These features are available through the HTTP connector which is easier to troubleshoot (sniff) and already standardized. 2. The HTTP connector was somewhat neglected; to ensure that it is completely conformant needs more eyes, not fewer. More effort at AJP 1.x is less effort towards HTTP/1.1 conformance. (This is not only a developer issue, but speaks to how well exercised the HTTP connector is with many users choosing AJP and not seeing or reporting specific quirks.) 3. We would honestly win more bandwidth from fully supporting the content encoding deflate from tomcat to the proxy server than from the few bytes saved with AJP. And SSL Encryption + deflate provided by TLS today will already give you this win, so binary protocol is really not that significant (OpenSSL 0.9.8 supports it, don't ask me if JSSE does.) 4. Waka. Why reinvent a wheel in motion? With the new focus at the httpd Amsterdam code to break apart http from apache, we are adding wiggle room for some to come behind and code to Roy's binary http protocol plan. The difference? Waka when done will be an accepted spec, while I don't see that ever happening to AJP. :) That said, those are technical arguments against but I have no vote here - I'll leave it to you all to weight these against your itches and designs. Bill --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
