On 8 July 2014 08:04, Roberto Peon <grmocg@gmail.com> wrote: > If you need to know the entire size of the headers before one is allowed > to send any of it (as would be the case), then one must wait for all of the > headers data to arrive (and mutate it) before forwarding. > Thus, adding at a minimum: header-bytes/bandwidth seconds of latency per > gateway (not to mention the additional memory for the buffering). > Hey this is double jeopardy! When I've previously pointed out the problem of DoS deadlock by sending a HEADERs and never sending the promised CONTINUATION, the response has been that no intermediary should every forward a HEADERs until it has the entire header block. So it can't be argued both ways - either we come up with something that truly allows streaming headers (ie no mutating of shared state) OR we accept that all headers have to be buffered. I'd expect this to add a 10th of a ms per gateway or so, more on more > constrained links. > only for the 0.001 of traffic that has large headers. Cry me a river! cheers -- Greg Wilkins <gregw@intalio.com> http://eclipse.org/jetty HTTP, SPDY, Websocket server and client that scales http://www.webtide.com advice and support for jetty and cometd.
Received on Monday, 7 July 2014 22:13:31 UTC