[rrg] Rejecting all but Strategy A

[Dino Farinacci]

I'm not seeing any compelling benefit here Pekka. Just that HIP is  
different and the stuff it does have that LISP currently doesn't is  
too heavy-weight. This is just my opinion and a judgement.

Dino

On Jan 7, 2009, at 3:55 AM, Pekka Nikander wrote:

> Eric wrote:
>>> By combining a true locator-identifier split (i.e., HIP) with a
>>> map-and-encaps (e.g., LISP) one gets the combined benefits of both
>>> approaches. Those benefits are
>
> Dino replied:
>>
>> But you add another layer of addressing that is probably more than  
>> the user wants. It could get this bad:
>>
>> HIT -> EID -> private-RLOC -> public-RLOC    when behind a NAT
>>
>> or
>>
>> HIT -> private-EID -> public-EID -> RLOC     when behind a NAT
>
> I don't think those mappings are right.
>
> If there is a NAT between the xTR/HIP-proxy and the "new" internet,  
> the mapping would be:
>
> EID -> (HIT) -> private-RLOC -> public-RLOC
>
> where the HIT would be invisible everywhere but inside the HIP  
> proxies and within the signalling protocol between the HIP proxies.
>
> If there is a NAT between the legacy host and the xTR/HIP-proxy, the  
> mapping would be:
>
> private-EID -> public-EID -> (HIT) -> RLOC
>
> where again the HIT would be invisible, as explained above.
>
>> 2-levels of mapping is plenty, adding 2 more probably is a non- 
>> starter. Remember folks, we have to incrementally deploy this with  
>> minimal cost, see Yakov's post early today, he makes good points.
>
> I don't see how replacing the xTR internal functionality with the  
> HIP-proxy functionality would add much cost.  It would change the  
> protocol between the xTRs, yes.  HIP as a protocol is probably more  
> complex than the present xTR-xTR protocol, but it is well tested,  
> been in operational use at Boeing for a few years, as has three open  
> source implementations.
>
> There is an extra mapping, i.e. replacing the EID->RLOC mapping with  
> a EID->(HIT)->RLOC mapping.  But that mapping is internal to the  
> proxies and the signalling protocol, and not visible outside.  In  
> the simplest case the HITs would be ephemeral and opportunistic,  
> requiring no configuration or storage anywhere.  (The benefit from  
> them is in terms of security and mobility, especially in more  
> complex scenarios than the simplest case.)
>
> --Pekka
>