Re: [Doh] WG Review: DNS Over HTTPS (doh)

[Mark Nottingham <mnot@xxxxxxxx>]

I suppose, although it would be good to understand if perf problems to a dhcp configured resolver are widely encountered. 

I wouldn’t be against text in the charter that allows the wg to consider mirroring currently defined (by the ietf) dns discovery mechanisms. I’d be against definition of new ones. 

Sent from my iPhone

On 22 Sep 2017, at 3:40 am, Phillip Hallam-Baker <phill@xxxxxxxxxxxxxxx> wrote:

On Thu, Sep 21, 2017 at 4:34 AM, Ask Bjørn Hansen <ask@xxxxxxxxxxxxxx> wrote:

> On Sep 16, 2017, at 9:07, Phillip Hallam-Baker <phill@xxxxxxxxxxxxxxx> wrote:
>
> 1) I see no evidence that HTTP/2 is suited to Web Services or will be dominant in that role. HTTP/2 was designed to serve Web Browsing to the exclusion of all other concerns. Which was the right choice to make.

HTTP/2 is also better for services with many small requests, in particular on high latency connections (or where each response might be slow to start…).

​If Web Services actually used HTTP features other than Firewall bypass and framing of transactions, then HTTP/2 might be attractive. ​Given how little of the HTTP stack is used and given that QUIC is a much closer match, that is the route I want to take.

I think it likely QUIC will eat up COAP as well. 

Just think of QUIC a way of doing TCP/2 in a way that is compatible with the protocol stacks as deployed in the field. At some point there will be a way to specify the service endpoint in a consistent fashion.