Eddie, That wording was unclear. I attempted to clarify in my later emails, but maybe I didn't succeed. I'm well aware that you need to consider the UDP ports for a 6-tuple receiver. As you have reminded us, you need this for NAT traversal. I think you also need this in order to make an effective user-space implementation of the encapsulation daemon, although clearly I'm being unable to explain that today.
tl;dr: I vote for the 6-tuple. Colin On 7 Feb 2011, at 16:36, Eddie Kohler wrote: > OK, I still don't get it but this is my last email on it. > > You said "the receiver operation is just to remove the UDP header, and treat > the encapsulated DCCP packet as any other native DCCP packet received. I'd > expect this also to be simple to implement as a user-space daemon." > > This does not describe the operation of a 6-tuple receiver, since the > described DCCP implementation does not consider UDP ports when looking up > flows. It describes the operation of an IP-plus-DCCP-ports receiver. > > Obviously there is some syntactic problem here, in my brain or in the wording. > > Eddie > > > On 02/07/2011 08:29 AM, Colin Perkins wrote: >> I'm well aware of the example. It's one of the reasons why I'm arguing for >> the 6-
