Re: Fwd: New Version Notification for draft-wouters-edns-tcp-keepalive-00.txt (fwd)

Mark Andrews <[email protected]> wrote:
>
> All DNS servers should already handle multiple questions over a TCP
> socket.  This is *basic* RFC 1035.  Just send the queries and deal
> with the fact that you may have broken servers.
>
> Adding a new option often the worst thing you can do to fix a failure
> to follow specification.  It is definitely not needed to fix this
> problem.

Right.

I have been thinking about what changes this draft implies relative to RFC
5966. There's a general encouragement to use TCP, but that doesn't require
protocol changes.

I don't think it implies any changes to server connection management.
There are quality-of-implementation questions, but they doesn't need
protocol changes.

The only difference it makes to the client is it can close connections
based on the server's idle timeout, instead of keeping them open
indefinitely.

This is beneficial to the server, since it offloads FIN-WAIT-2 to the
client. It is not beneficial to the client since it implies extra timeout
handling and extra bandwidth usage - think radio power.

Perhaps the server should only send the option when its timeout is less
than indefinite, or if it changes in the lifetime of the connection. Then
in most cases the client will not have to bear the extra costs.

Tony.
-- 
f.anthony.n.finch  <[email protected]>  http://dotat.at/
Forties, Cromarty: East, veering southeast, 4 or 5, occasionally 6 at first.
Rough, becoming slight or moderate. Showers, rain at first. Moderate or good,
occasionally poor at first.
_______________________________________________
dnsext mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsext