Re: Question re HIP dependency [Re: about Wedgelayer 3.5 / Fat IP approaches]

On 25-jul-04, at 20:36, Jari Arkko wrote:

> I also think that posing the question at this time is a bit unfair on 
> HIP. The responses that you got were that its either OK to use HIP or 
> that HIP lacks feature X or Y which make it problematic to use it. The 
> problem is, most if not all proposals in the category of solutions we 
> are considering have such deficiencies; our task is to compose or 
> develop a solution which overcomes these difficulties.

The way I understood Brian's question is: would it be ok to build on 
HIP the same way some proposed solutions build on the DNS, ie., just 
assume that it's there and that it works.

My answer to this was that there are two issues: the first is that HIP 
isn't deployed yet. This will/can be solved in time and propbably isn't 
all that important in the long run. Still, building on HIP _now_ would 
probably make our life more difficult. The second is that HIP as it is 
today has certain properties that are undesirable for a generic 
multihoming solution. (Larger packets and mandatory crypto.) This would 
be like using the DNS but the DNS takes 15 minutes to resolve a query: 
you can build such a solution, but it doesn't address the user needs 
very well..

So basically what we're saying is that we only want to use HIP if it's 
modified to our liking.  :-)  This isn't unfair as we want all the 
solutions to be modified to our liking.  (-: