RE: Why is PMIP a superior solution for NETLMM requireiments?(was:Re: [netlmm] Network-based mobility - NETLMMDT SolutionVsAsolutionbased on MIP6)

At the last couple of IETFs, we have discussed this quite a bit, but the
topic seems to be coming up again. In my view, the use of network-based
mobility for anything other than local mobility is bad. Global mobility
management is meant to deal with mobility across heterogeneous
interfaces and mobility across administrative domains. Network-based
mobility solutions have inherent problems with such scenarios. Just
because PMIPv6 is based on MIP6 does not mean that it automatically has
the same scope as MIP6. IMHO, NETLMM and PMIP6 are solving the same
problem and suffer from exactly the same limitations in terms of
applicability. 

As many would recall, we had a discussion in Dallas on network-based
global mobility and talked about why it is a bad idea. As far as I know,
nothing has happened to change that view. 

Regards,
Vidya

> -----Original Message-----
> From: Vijay Devarapalli [mailto:vijay.devarapalli@azairenet.com] 
> Sent: Monday, October 30, 2006 9:04 PM
> To: Templin, Fred L
> Cc: netlmm@ietf.org
> Subject: Re: Why is PMIP a superior solution for NETLMM 
> requireiments?(was:Re: [netlmm] Network-based mobility - 
> NETLMMDT SolutionVsAsolutionbased on MIP6)
> 
> Hi Fred,
> 
> Templin, Fred L wrote:
> > Having glanced through 'draft-sgundave-mip6-proxymip6', 
> I'll ask the 
> > obvious question: isn't it true that the PMIP proposals 
> (or, at least 
> > the sgundave proposal) are about maintaining a stable 
> *home* address 
> > while the NETLMM/DT and NETLMM/DHCP proposals are about maintaing 
> > stable *regional* addresses? If so, then is it possible 
> that the two 
> > solutions are complementary rather than one vs. the other?
> > 
> > For my application, the stable regional address is probably 
> going to 
> > best satisfy the majority expected use cases since traffic 
> won't have 
> > to flow through a (possibly) distant home agent. But it seems like 
> > other applications could benefit from a stable home address 
> since it 
> > would give the MN something stable to put into the global 
> DNS whereas 
> > a stable regional address alone might not.
> 
> this makes sense when you do use a global mobility management 
> protocol *and* a local mobility management protocol at the 
> same time. the equivalent when you use host-based mobility 
> management protocols is running MIPv6 and HMIPv6 at the same time.
> 
> but there are many deployments where there are no two levels 
> of IP mobility. there are also deployments where there are 
> two levels of IP mobility, but the local mobility management 
> protocol is at L2.
> 
> when you consider a particular domain which is running just 
> *one* network mobility management protocol, there is no 
> difference between running PMIPv6 or NETLMM. MAG is 
> equivalent to the PMA, LMA is equivalent to the HA and the 
> regional IP address/prefix is the same as the home 
> address/home prefix.
> 
> as far as I am concerned, we are not labeling PMIPv6 as a 
> network-based global mobility management protocol. it can be 
> used a network-based local mobility management protocol. in 
> fact I think its a bad idea to use anything other than a 
> host-based mobility management protocol when the mobility 
> signaling is happening over the Internet.
> 
> Vijay
> 
> _______________________________________________
> netlmm mailing list
> netlmm@ietf.org
> https://www1.ietf.org/mailman/listinfo/netlmm
> 

_______________________________________________
netlmm mailing list
netlmm@ietf.org
https://www1.ietf.org/mailman/listinfo/netlmm