Re: [Mip4] RE: Differences between Low Latency Handovers and FMIP v4

[Charlie P <charliep at iprg.nokia.com>]

Hello Kent,

Kent Leung wrote:

> My personal take on this is that the early agent discovery is unnecessary
> in the overall latency.  The return on investment of the new messaging 
> and
> operation is minimal.  I think having the agent advertisement sent by the
> new FA after access authorization will provide fast handover.  The 
> time for
> MN to learn about FA can be sub-msec range.

This would be a specification written for very particular systems that
had the hardware characteristics you see.  Other people have been
persuaded that the only possible solution was make-before-break.
I am much in favor of a more widely applicable layer-3 solution,
particularly because it enables better performance in systems with
multiple physical media available for Internet connectivity.


>   With PFANE and BU/BA,
> the packets in transit are no lost during registration period to acheive
> the smooth handover.

I don't think this is correct.  But I will agree that the motivation for
PFANE was to minimize packet loss (and of course eliminate if it
happens to be a good day).

>   It is my opinion that we can achieve good optimization
> with some basic messages and extension, without adding much complexity
> to the mobility agents.

We should structure the solution so that, in the case you envision,
the minimal implementation is enabled by the specification.

Regards,
Charlie P.

--
Mip4 mailing list: Mip4 at ietf.org
   Web interface: https://www1.ietf.org/mailman/listinfo/mip4
    Charter page: http://www.ietf.org/html.charters/mip4-charter.html
Supplemental site: http://www.mip4.org/