RE: [netlmm] #55: Define desired scalability more precisely

["Narayanan, Vidya" <[e-mail]>]

Hi Pekka,

<snip>

> 
> I'd rather state explicitly that the solution must be able to 
> scale to, say, O(10K) or O(50K) simultaneous users.  

I can't understand the reasoning behind explicitly stating any number
here. Any numbers for scalability are system level requirements and not
protocol level requirements. The only goal of a protocol should be that
it be scalable - in the sense that in a real deployment, it must be
possible to *scale up/out* the protocol. For e.g., if the NETLMM
protocol had limitations in the number of LMAs that can be present for
the protocol to function properly, that would be a bad protocol design.
But, as long as more users can be supported by increasing the number of
LMAs and/or ARs, the protocol itself is scalable. System level
scalability requirements can then say x number of simultaneous users (to
figure out how many actual boxes are needed) or x number of users per
LMA, etc., but this is outside the scope of the protocol itself. 

> If the 
> implementations can work on the scalability to make that much 
> higher -- that's great, but not something we in the IETF 
> should be concerned about, because networks can just deploy 
> more boxes without significant disadvantage to the user if 
> the limit is reached.
> 

As I said above, I don't think the numbers mean anything to a protocol
and should not be part of any protocol requirements. 

> It seems to me that as soon as we discard Mobile IP as a 
> basis for localized mobility management, there seems to be 
> push to make the LMM domain bigger and bigger so that you 
> could make do without any non-LMM mobility management 
> solution.  As shown in the list a)-f), building one LMA/MAP 
> big enough to cover the whole world (or break users' 
> session if (s)he moves e.g. from one city to another) just 
> doesn't seem feasible, so I can't figure out why we're trying.
> 

This itself is a valid concern and I do believe that an applicability
statement or the like in the document must be added to clarify that
there are limitations in making the LMM domain bigger and bigger. 

Regards,
Vidya

> -- 
> Pekka Savola                 "You each name yourselves king, yet the
> Netcore Oy                    kingdom bleeds."
> Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
> 
> _______________________________________________
> netlmm mailing list
> [e-mail]
> https://www1.ietf.org/mailman/listinfo/netlmm
> 

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