Re: [IPsec] Informal poll on IKEv2 { over TCP | fragmentation }

[Yoav Nir <[e-mail]>]

On Mar 14, 2013, at 7:33 AM, Paul Hoffman <[e-mail]> wrote:

> <chair-hat on>
> 
> It is seeming that consensus is trending towards "let's document the fragmentation method some vendors are already doing instead of finishing IKEv2-over-TCP", but I would like to check that. Please respond to the informal poll below.
> 
> --Paul Hoffman
> 
> 
> - I would prefer the WG to continue working on IKEv2 over TCP
> 
> - I would prefer the WG to stop working on IKEv2 over TCP and instead work on standardizing IKEv2 fragmentation
> 
> - I would prefer the WG to continue working on IKEv2 over TCP and also work on standardizing IKEv2 fragmentation

In another meeting, I said that standardizing two ways of achieving the same goal is a failure mode. I still believe that. I also believe that fragmentation is easier to fall back on than TCP. IOW an IKE-over-TCP draft would likely become "always use IKE over TCP" in a product (my company's clients that had IKE-over-TCP did exactly that), whereas fragmentation is more likely to become "use fragmentation if the the regular UDP did not get an answer". So:

"I would prefer the WG to stop working on IKEv2 over TCP and instead work on standardizing IKEv2 fragmentation"

However, I would prefer that we clear up the situation with Microsoft's IPR before making such a change to our charter.

Yoav