Re: [MEXT] [dmm?] Surprising assertion about make-before-break ...

[Hesham Soliman <[e-mail]>]

Hi Charlie,


>
>Hello Hesham,
>
>Long time no see...

=> Likewise! 

>To your point:
>
>On 8/4/2011 10:11 PM, Hesham Soliman wrote:
>
>> On the battery issue, yes of course but a few years ago those smart
>>phones
>> needed to be charged every few hours, now they improved. My laptop
>>battery
>> lasts 5-6 hours, which is an improvement. So these things change over
>> time.
>
>Assume, just for a moment, there were a handover method requiring
>only a single radio interface to be powered up at any one time, that
>had roughly equal performance as handover algorithms which relied on
>multiple radio interfaces being powered up.
>
>In that case, I think it would be considerably preferable to
>make use of the single-radio algorithm.

=> Sure. I think we got two topics wrapped up in the same thread, I was
referring to multiple radio interfaces using different technologies. But I
agree with you that you can get good performance without make before break
handover. 

>
>Excellent performance with single-radio devices is quite possible,
>and we would practically already be there except for politics or,
>(less likely but plausible) simply a matter of inertia.  Certainly
>possible for good VoIP, almost certainly for interactive video on
>4G networks <--> WLAN.  We were showing smooth VoIP handovers
>almost ten years ago with 802.11b, and with SFF-based preregistration
>in 4G networks we could do far better than that now (I recently
>submitted a draft about this).

=> Absolutely, I've done similar experiments on WLAN and still do in fact
and I agree. And did the same with 3G and you're right, it's all politics
that's stopping this. Luckily for me I'm not in the middle of these
politics anymore :)

>
>How to get it standardized in LTE seems to be a puzzle of monumental
>proportions.  Perhaps if the end users realized how much better
>their service could be, and started demanding it, things would
>progress.  In the meantime, they'll get slow, battery-wasting
>handovers to WLAN that do not even preserve IP addresses, much
>less offer session continuity.  And the operators will have to
>purchase unbelievably complicated equipment to even enable that
>level of service.

=> Forget about average users demanding proper IP networks, some operators
tried that and failed. Unfortunately the "getting it into LTE" part is not
fun. The fun part is making it work and the rest is...well you know. But
it might be worth giving LTE a unified IETF solution for mobility instead
of re-dressing GTP in IP (PMIP) and then getting it rejected anyway. I
didn't like that approach, but that was a long time ago.


>
>I'd like to see the IETF start offering solutions that
>have an easier evolutionary path towards deployment.
>http://www.psg.com/~charliep/txt/ietf81/alt_mext/MIPv6_for_4G.pptx

=> Nice plug :), but I don't see the relation to the single Vs multiple
radio discussion.

Hesham

>
>[note I renamed that file, there was a typo in the previous name]
>
>Regards,
>Charlie P.