NNSquad - Network Neutrality Squad

NNSquad Home Page

NNSquad Mailing List Information

 


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ NNSquad ] Re: More on UMA - Unlicensed Mobile Access


It's not really a question of mythologies, Bob. Engineers have measurement tools that rate voice quality according to the MOS scale and we can prove that QoS over Wi-Fi increases the capacity of a given AP in a typical office scenario to carry calls at a given level of quality. Measurements indicate that QoS-enabled Wi-Fi can carry 4 times as many calls as best efforts Wi-Fi, all other factors being equal and typical. The Wi-Fi standards engineers who wrote 802.11e don't appreciate complexity for its own sake, but they do like systems that work well.

RB

On 1/23/2011 6:48 PM, Bob Frankston wrote:
Saying that it works because of QoS is presumptuous for 8kbps on a 1000kbps
link. So, at best, we're talking about a single wireless link. Commercial
VoIP doesn’t depend on QoS for the entire path.

I have used Skype very well over below-quality long latency connections and
it does remarkably well by taking advantage of what is available. It's like
the Mocalliance claim that IP was not designed for video even as Verizon
runs VoD over IP.

We can see the harm done by this assumption in the problems with FiOS. Even
though video worked well through my private router and through my current
router it runs into a problem with the new one because of a problem with the
STUN protocol. But Verizon has lashed its future to the user of their
router. Thus their applications like remote DVR only work only if you use
their routers. This puts the kibosh on innovation.

The use of protocols with QoS doesn't mean it is necessary and, in fact, you
can't assume QoS beyond the local link. I do know that the application on my
phone checks to see the actual capability of the path rather than trusting
the specifications.

We also saw this in Firewire (IEEE-1394) which failed because there was too
much application knowledge in the protocol.

So let's not spread mythologies -- just because a particular protocol exists
doesn't mean it is necessary and making it necessary limits our ability to
move beyond the narrow vision of the initial specifications.


-----Original Message----- From: nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org [mailto:nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org] On Behalf Of Richard Bennett Sent: Sunday, January 23, 2011 20:28 To: Bob Frankston Cc: nnsquad@nnsquad.org; dewayne-net@warpspeed.com; 'Esme Vos' Subject: [ NNSquad ] Re: More on UMA - Unlicensed Mobile Access

UMA works on the unlicensed Wi-Fi spectrum quite well, in my experience,
and it does so for a couple of reasons. For one, Wi-Fi has explicit QoS
support since 802.11e (sometimes called WMM) that good UMA
implementations know how to use; and for another, Wi-Fi operating
indoors and at low power - especially 802.11a - doesn't have to contend
with a great deal of interference from other networks and users. That
doesn't mean that you can scale Wi-Fi into a high-power, outdoor network
and expect the same functionality.

But you're welcome to try, as many people have.

RB

On 1/23/2011 5:12 PM, Bob Frankston wrote:
For more on UMA


http://www.eetimes.com/design/signal-processing-dsp/4016283/UMA-Demystified-
Inside-UMA-enabled-dual-mode-handsets

It has interesting tidbits including a mention of APSD - Automatic Power
Save Delivery in 802.11e which provides some degree of power management. I
don't know how available it is.

As I've assumed in the past with the earlier T-Mobile UMA handsets you
can't
just add it since it's in the "FCC cage" but apparently the ability is
latent in a number of phones such as the Vibrant.

There's obviously lots more on the net including

http://kantarainitiative.org/confluence/download/attachments/37751312/UMA-pr
otocol-flow-deep-dive.pdf which gives the protocol exchanges.

UMA liberates us from the physical artifacts of cellular but not the
business model. There are also annoying legacies like apps that need
explicit 3G for security.

What is important is that UMA demonstrates that wireless VoIP is
mainstream.
It's interesting that the main feature seems to be "unlicensed" and,
indeed,
licensing is central to today's telecommunications industry. What would
happen if we didn't have to ask for permission to speak?


-----Original Message----- From: nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org [mailto:nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org] On Behalf
Of
Richard Bennett
Sent: Saturday, January 22, 2011 19:05
To: Bob Frankston
Cc: nnsquad@nnsquad.org; dewayne-net@warpspeed.com; 'Esme Vos'
Subject: [ NNSquad ] Re: Wi-Fi Calling!

T-Mobile had Wi-Fi based UMA (Universal Mobile Access) calling before
Android, I used it on from a Blackberry on TM back in 2007. Their
billing model is interesting: if you start a call on Wi-Fi and then get
in your car and drive off, UMA will hand the call over to cellular but
you won't be charged any minutes. But if you start a call on cellular
and UMA hands it over to Wi-Fi, you'll be charged minutes for the whole
call.

Or so it used to be, I stopped using TM after the G1 became their only
3G smartphone and I found it inadequate. Given that the call has to be
connected through the TM network to some other party, their costs don't
go away simply because one side of the route to and from their network
is paid for by some other operator.

RB

On 1/22/2011 3:34 PM, Bob Frankston wrote:
     [ This is not a Samsung-specific feature.  T-Mobile reportedly
       rolled out similar functionality for a variety of their
       high-end smartphones (including the HTC-built G2) a couple of
       months ago I believe.  I've been meaning to look into the
       security of the protocols being used.

             -- Lauren Weinstein
                NNSquad Moderator ]


The T-Mobile Samsung Vibrant upgrade has a very interesting feature = =E2=80=93 a Wi-Fi calling application. This means you can use your = T-Mobile account even if you don=E2=80=99t have T-Mobile coverage as = long as you get Wi-Fi access. This dramatically improves the ability to = make calls.

=20

Admittedly you=E2=80=99re warned that you will be using =E2=80=9Cplan =
minutes=E2=80=9D which is a separate policy issue. Almost by accident it
=
renders femto cells moot. For those traveling in Paris near a park for =
example you can use the available Wi-Fi coverage. The application allows
=
you to set preference for cellular vs. Wi-Fi as the preferred means.

=20

The upgrade itself is interesting =E2=80=93 I only found it by searching
=
the web rather than navigating from the T-Mobile home screen and I only =
know about it from having a helpful store manager.

=20


http://forums.t-mobile.com/t5/Samsung-Vibrant/Samsung-Vibrant-software-up=
grade-to-Android-2-2-Froyo-now/td-p/678871

=20

You can=E2=80=99t do it over the air =E2=80=93 you must use their Kies =
software which, despite Samsung=E2=80=99s claims, does not work on =
Windows 7. But I did scrounge up an XP machine and on the second try got
=
it to work =E2=80=93 hint =E2=80=93 do a hard reboot first.

=20

Why not OTA? I=E2=80=99d like to know. Is T-Mobile charging Samsung for =
=E2=80=9Cair minutes=E2=80=9D. With Wi-Fi calling you=E2=80=99d think =
that they would want people to upgrade their phones. Perhaps =
it=E2=80=99s nothing nefarious and the fact that I had to do a hard =
reboot is part of the issue of Samsung vs. software. I have software =
upgrade problems on my Samsung TV too and each model has a different =
version of the software (see http://rmf.vc/ces2011.nn for more on this).

--
Richard Bennett


--
Richard Bennett



-- Richard Bennett