- From: Amos Jeffries <squid3@treenet.co.nz>
- Date: Mon, 01 Jul 2013 15:51:36 +1200
- To: ietf-http-wg@w3.org
On 1/07/2013 3:28 p.m., David Morris wrote: > > > On Sun, 30 Jun 2013, Michael Sweet wrote: > >> I haven't had a proper amount of time to review the current draft, but I just wanted to throw something out... >> >> >> On 2013-06-29, at 5:14 AM, Yoav Nir <ynir@checkpoint.com> wrote: >>> On Jun 29, 2013, at 9:04 AM, David Morris <dwm@xpasc.com> wrote: >>> >>>> I continue to suspect that PUSH_PROMISE is either not going to be >>>> heavily used on the general web or it is going to waste significant >>>> bandwidth and server capacity. >>> There's been very little talk (at list on-list) about how a server decides what resources to push. One way is to push down all the resources referenced from an HTML document. That would interact badly with caching, both in clients and in proxies. >> Currently a client will send GET requests with If-Modified-Since for every resource it needs; perhaps if there was a way for the client to pass along a list of already-fetched resources with the fetch date, the server could push any new resources back to the client? >> >> (that might defeat any performance advantage, or maybe not - something to try?) > I think the basic premis is that PUSH_PROMISE is generated by the server > before the client has had a chance to receive and parse the original > request and identify what it has and might need. There does seem to be an interesting case for RST_STREAM+If-Not-Modified here to optimize bandwidth usage. Amos
Received on Monday, 1 July 2013 03:52:06 UTC