- From: Amos Jeffries <squid3@treenet.co.nz>
- Date: Tue, 10 Dec 2013 10:31:14 +1300
- To: ietf-http-wg@w3.org
On 2013-12-10 05:43, Martin Thomson wrote: > On 9 December 2013 01:18, Amos Jeffries wrote: >> LadleNet > > I'm not sure what to make of this one. It sounds a lot like the > Internet. > Does doesn't it? if the protocol can't support *that*, well ... But no its not. Just the microcosm of the whole in one of many Ad-Hoc mobile environments. The key parts for HTTP to operate over this environment are: * message transparency to the node proxy classification filtering regardless of encryption or other security overlays on the application payload. * HTTPS encryption is not an option both because it breaks the node classification and prevents fast per-message re-routing when edge/gateway nodes disappear or move. * "Just use HTTP/1" is the worst of undesirables. Limited/slow bandwidth and CPU already pressures heavy erasure of HTTP/1 meta data and content compression at the edges. > mnot wrote: >> Not every problem needs to be hit with an HTTP-shaped hammer. Maybe not. But HTTP is used to traverse this one and for certain types of inter-node exchanges there are limitations to be met. Amos
Received on Monday, 9 December 2013 21:31:40 UTC