One thing I think has always been bugging me about DNS over HTTP is the
appalling performance we will likely see in a lot of cases. Even small
latencies in normal DNS lookups cause major impact on page load times on
complex pages with resources from many servers, and adding HTTP overhead
to this will slow things further.
So I suggest some thought should be given to reducing round-trips by
allowing for multiple DNS request payloads in a single HTTP request
message, and multiple DNS response payloads in an HTTP response message.
E.g. sending requests as base64 encoded blob values in key-value pairs
as post data, so that a response can be linked back to the request.
At this stage of the game this should be quite easy, but will become
more and more difficult as time goes by if the single request/response
model is rolled out.
Regards
Adrien
------ Original Message ------
From: "Tim Wicinski" <tjw.i...@gmail.com>
To: "dnsop" <dnsop@ietf.org>
Sent: 12/07/2016 10:33:14 a.m.
Subject: [DNSOP] Call for Adoption: draft-song-dns-wireformat-http
This starts an official Call for Adoption for
draft-song-dns-wireformat-http
The draft is available here:
https://datatracker.ietf.org/doc/draft-song-dns-wireformat-http/
Please review this draft to see if you think it is suitable for
adoption by DNSOP, and comments to the list, clearly stating your view.
Please also indicate if you are willing to contribute text, review,
etc.
We wanted this Call to coincide with the Berlin meeting so if there is
opinions that needed to be voiced, they can do so.
This call for adoption ends: 25 July 2016
Thanks,
tim wicinski
DNSOP co-chair
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
