On Mon, Jan 04, 2016 at 08:10:09AM -0800, internet-dra...@ietf.org <internet-dra...@ietf.org> wrote a message of 43 lines which said:
> Title : A review of implementation DNS over port 80/443 > Authors : Shane Kerr > Linjian Song > Runxia Wan > Filename : draft-shane-review-dns-over-http-02.txt 2.1 "DNS over TCP on port 80/443" "One drawback of this approach is that it might mislead the client because of the port used." Another drawback for port 443, is that a zealous firewall may DPI the content, find out it is not TLS, and decide to drop/block. 2.2 "DNS over TLS on port 443" says "One concern of DNS over TLS is its cost. Compared to UDP, DNS-over- TCP requires an additional round-trip-time (RTT) of latency to establish a TCP connection." No longer true if 5966bis is implemented and deployed. 2.3 "DNS Wire-format over HTTP(S)" contains "Using HTTP also benefits from HTTP's persistent TCP connection pool concept (see section 6.3 in [RFC7230]), which DNS on TCP port 53 does not have" Same remark as the previous paragraph, you should mention 5966bis. Editorial : > It is quiet useful quite _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
