Re: [P2PSIP] HIP: optional, mandatory?

[Gonzalo Camarillo <[e-mail]>]

Hi Henning,

> At least we're now clear that the HIP advocates want to essentially 
> force systems to implement it. This will hopefully focus the discussion 
> a bit.

not at all. I, for one, do not want to force anyone to implement HIP if 
he or she does not want to.

I agree with you that if you buy a product that supports "IETF P2PSIP" 
and I buy another product that also supports "IETF P2PSIP", they should 
be able to talk to each other... of course, we may find out that our 
boxes do not implement the right DHT, but that is a different story.

Of course, it would be OK for the P2PSIP WG to decide not to use HIP for 
"IETF P2PSIP" applications, if this is what most people want (this is 
what is being discussed). You are also right that, lately, we tend to 
build too complex systems and that we should be aplying the kiss 
principle more often... but I come from a different perspective.

I want to build a HIP-based overlay. If the P2PSIP WG decides that 
P2PSIP should be HIP-free, that's OK. If that was the case, I would run 
other types of applications on top of my HIP-based overlays.

The thing is that, in order to build my HIP-based overlay, I need a peer 
protocol whose functionality is a subset of the functionality provided 
by the peer protocols designed in the P2PSIP WG... therefore, in order 
to avoid reinventing the wheel once more, I would like to use those peer 
protocols for my HIP-based overlay.

Consequently, my requirement on the peer protocol is documented on 
Section 5 of the HIP BONE draft:

        ... the peer protocol
        standardized by the WG is kept functionally and specification
        wise reasonably modular so that the HIP community can use the
        peer protocol minus the connection management and NAT traversal
        modules to experiment with HIP-based overlays.  Note that, at
        present, this is the case with several peer protocols.

I really hope nobody has a problem with this requirement.

With respect to whether or not option 3 (optional to implement and run) 
is actually an option, we can discuss whether or not making HIP optional 
to run for any given node in any given overlay adds too much complexity 
to a system (the capability negotiation mechanism needed for that and 
the stuff that would need to be implemented twice). I personally tend to 
think that the complexity would be relatively high, but some folks do 
not think so.

Cheers,

Gonzalo

_______________________________________________
P2PSIP mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/p2psip