- From: Jason Greene <jason.greene@redhat.com>
- Date: Thu, 19 Jun 2014 20:23:04 -0500
- To: Poul-Henning Kamp <phk@phk.freebsd.dk>
- Cc: Martin Thomson <martin.thomson@gmail.com>, Mark Nottingham <mnot@mnot.net>, HTTP Working Group <ietf-http-wg@w3.org>
On Jun 18, 2014, at 4:21 AM, Poul-Henning Kamp <phk@phk.freebsd.dk> wrote: > In message <CABkgnnUeS823DWFY+CdBaKF2GUzc4_QJoRf1RXVcqxgKhN-f7A@mail.gmail.com> > , Martin Thomson writes: > >>>> page 57: query part >>>> ------------------- >>>> >>>> The query part of the URI should get its own ":query" psuedo-field >>>> for reasons of othogonality and to reduce the amount of data typical >>>> load-balancers have to examine. >> >> I think that we discussed this and decided against it. Happy to >> follow up, but I'd have thought that the gains would be marginal. > > Was that discussion before or after HPACK ? > > I would expect HPACKs efficiency to increase if the constant-ish > :path part were divorced from the dynamic-ish :query part ? I think you are on to something with separating headers which represent sensitive user data from the more transport oriented data. If it’s separate it can potentially be encrypted without the need (or desire) for the proxy to decrypt it (e.g. the bank example in the strawman thread). -- Jason T. Greene WildFly Lead / JBoss EAP Platform Architect JBoss, a division of Red Hat
Received on Friday, 20 June 2014 01:23:44 UTC