AT&T's Foggy FMC Plans

There's a little bit of mystery surrounding AT&T Inc. (NYSE: T)'s launch of the BlackBerry 8820 WiFi-enabled smartphone, due out late this summer.

BlackBerry made some play of the fixed/mobile convergence (FMC) capabilities of the new phone when it launched Tuesday afternoon. The Canadian device maker noted that the dualmode device supports the unlicensed mobile access (UMA) technology, pioneered by startup Kineto Wireless Inc. With a UMA system in place, the BlackBerry 8820 can switch voice calls between a wireless carrier's cellular network and a WiFi network, according to RIM. (See RIM Goes WiFi.)

AT&T, however, won't be drawn on whether it will use this capability to offer FMC services for 8820 users. "The only thing we're acknowledging at the moment is that we're launching it later this summer," an AT&T spokesman that deals with the devices tells Unstrung.

Our calls for a more general comment on the state of AT&T's overall FMC strategy have so far been unanswered.

In March 2006, AT&T -- then Cingular -- said that it would leapfrog UMA deployments and instead install an IP Multimedia Subsystem (IMS). The operator had initially undertaken several trials of UMA-based systems in 2005. Nothing, however, apparently came of those at the time. In January 2007, the carrier said that it had deployed IMS and was working on FMC but gave no more information. (See Cingular: The Call of the WiFi.)

Since then the operator has provided no firm dates for any kind of FMC launch. In the meantime, its main GSM rival, T-Mobile US Inc. , has introduced the first FMC service in the U.S. -- [email protected] -- using unlicensed mobile access technology.

AT&T's FMC strategy may become clearer as more details on its home base-station requirements emerge, although, so far, analysts don't have a lot to say on exactly what AT&T desires from a femtocell. (See Is AT&T Putting Out Femto Feelers?)

What they're looking for isn't really "ground-breaking" and is "fairly basic," according to Allied Business Intelligence Inc. (ABI) analyst Stuart Carlaw of AT&T's rumored femtocell RFP. "It's really about coverage and call plans… They're not looking to do a huge amount of multimedia." — Dan Jones, Site Editor, Unstrung

Page 1 / 2   >   >>
paolo.franzoi 12/5/2012 | 3:05:14 PM
re: AT&T's Foggy FMC Plans
I have been thinking about FMC for my own personal use (and may visit a T-mobile store in the near future).

What I have read so far:

1 - The service costs more than just plain mobile service. That makes no sense to me, but that is my understanding.

2 - They force some particular routers down your throat, but I also understand that you can toss their routers in the garbage and use your own.

Anybody out there have the service and have any commentary?

wap545 12/5/2012 | 3:05:13 PM
re: AT&T's Foggy FMC Plans Do we know if AT&T is restricting use of the WiFi mode on the 8820 in anyway or is it ignoring it.
Interesting that they have allowed their last two big handset releases to have access to and use of the WiFi feature which will dominate this entire Broadband Data space over the next few years.
The fact that this product has UMA capabilities, not sure what that means, is meanigless without a Controller interface.
Without a UMA or IMS based Controller at the Cell Network as an interface between the IP based WiFi network and the Cell Net nothing will happen as far a hot handoffs of call (VOice) or data between these two networks.
You will be able to make VoiceIP calls at a HotSpot or on the new WiFi based Metro Mesh Networks-as long as you have a VoiceIP service but not hand of the call to a Cell network.
All CellCO are afraid of the emergence of a very cost effective VoiceIP solution that will eventually imapct their big revenue source VOice and effectively sidetrack their Narrowband Data Networks-especially the CDMA based EV-DO nets.
This is why these CellCo have to win a piece of the nationwide 700Mhz spectrum coming up for auction in Feb 2008-they (mainly Verizon Wireless) need to keep competition out and enhance their data services.

paolo.franzoi 12/5/2012 | 3:05:07 PM
re: AT&T's Foggy FMC Plans
Okay, well I will answer my own questions after a visit to the T-Mobile store.

1 - T-Mobile offers the same plans as is available with your regular phones. There is one additional option and that is $10/month for unlimited WiFi calls. Otherwise they count out of your regular minutes.

2 - I could use my own router. On top of that, I can theoretically use any open network and any T-Mobile Hotspot. There is no additional charge to use a Hotspot.

I have not found that anything switches automatically. I have to tell the phone to use WiFi up to now, but will continue to look for more options.

It has greatly improved my coverage at home and am greatly pondering dumping my wireline. Setup was trivial and the phone (I picked the Samsung lower end phone) works fine.

If anybody has any questions, just post away.

IPobserver 12/5/2012 | 3:05:06 PM
re: AT&T's Foggy FMC Plans Hi brookseven,

Thanks for the info.

How does the hand-over from WLAN to cellular work for you? That's supposed to be automatic.

Would you have opted for a flashier phone if one had been available?
paolo.franzoi 12/5/2012 | 3:05:05 PM
re: AT&T's Foggy FMC Plans

I have not seen automatic handover from Cellular to WiFi. I have seen it the other way around and it was completely seamless. In fact, I did not notice it until I looked at my phone and noticed I was on Cellular. I was on a conference call and walked into my backyard. I went to unmute to say something and noticed I was on Cellular. I called in on WiFi, so that seamless operation was impressive.

I will have to pay better attention to the switch to WiFi, but I am NOT clear that this is automatic.

There was a flashier phone available, with a memory stick and a media player. I felt no need to take that phone, but it was the same price.

I really like the service so far.

jimgrams 12/5/2012 | 3:05:04 PM
re: AT&T's Foggy FMC Plans I have used the T-Mobile @ Home service in the Seattle area for quite some time. This service only automatically hands "out" from WiFi to Cellular, in my experience. It does this rather well, actually, not switching too soon, and only very rarely dropping the call. The hand "in" function is less important as my GSM coverage at home is okay, so if I start a call outside (such as in my car on the way home)I can complete it at home without needing the handover to Wi-Fi. If TMO GSM coverage stunk at my home, I'd feel differently, of course.

lrmobile_kampar 12/5/2012 | 3:05:02 PM
re: AT&T's Foggy FMC Plans
Apologies in advance for the long and somewhat technical answer ... hopefully this will demonstrate how this all works ... although some GSM knowledge might be required!

Handover to a UMA/GSM Dual Mode Handset is simpler than the femtocell case as there is a UMA client residing on the mobile device that can co-ordinate the handover with the network. In the femtocell case, you have to work with the handsets that are already out there, which presents a different set of (solvable) problems using UMA.

In dual-mode UMA applications, when the handset moves into WiFi coverage it detects the WiFi signal, checks to see if it has been associated previously with this Access Point, makes a WLAN connection, then registers with the UMA Controller.

If the handset is in idle mode during this process, it will then just 'rove-in' to the UMA Controller and appear as a GSM device connected to the core network through the UMA controller (just as if it was connecting through a 2G BSC).

If the handset is in dedicated mode (on a voice call) on the macro GSM network then the situation is slightly more complex.

First, the handset needs to know to look for an 'adjacent' cell to perform the handover to. In the pure GSM environment this is performed using the neighbour list table in the BSC. Each GSM cell has a list of known (provisioned) neighbours (the surrounding GSM cells) and the handset scans for these when it is on a call to see if it finds one that has a stronger signal than the current cell. If it sees a stronger signal, the handset notifies the BSC which kicks off a handover procedure to move the handset to the new GSM cell.

In the UMA case this is a problem. There could be hundreds of WiFi access points in a single GSM macro cell, each of which could be considered a neighbour to the macro cell for handover purposes, however the BSC can typically only configure relatively few neighbour cells per macro cell. Therefore you cannot (and nor would you want to) provision each and every WiFi access point as a neighbour to each cell in the macro network.

The answer is to abstract the provisioning of the UMA access network from that of the macro network. UMA Controllers typically do this by representing a large number of UMA end-points (handsets) as one or more GSM cells to the core network. In this way you can minimize the provisioning burden of the UMA neighbour cell tables in the macro network (each GSM cell has only one neighbour representing the entire UMA population within its cell area) which is obviously desirable for the operator.

The exact way in which each UMA vendor does this is often part of their IPR but the idea is that standard techniques are used in the macro network (i.e. no changes required to support UMA in the installed macro and core networks, vs. alternate methods of accomplishing handover using WiFi which often require significant modifications).

The piece that makes this work however is the client on the UMA handset. When the UMA client registers with the UMA Controller it receives back information about the 'abstracted' UMA environment (e.g. the UMA cell 'frequency' which doesn't really exist in the physical radio network).

If this happens during a call, then the handset now magically 'sees' the neighbour UMA cell frequency as represented in the neighbour table for the macro cell. This is because the UMA client on the handset forces this information into the GSM Radio Resource Management algorithm ... i.e. it 'forces' the GSM layer into believing that it has seen a very strong macro signal representing one of the neighbours in its scan list.

The handset then tells the BSC it has seen a strong signal from an adjacent cell, the BSC kicks off a handover request to the core network which is forwarded to the UMA Controller. The UMA Controller knows from the handover request which handset is attempting handover and it knows where that handset is in the IP access network as it is already registered to that same UMA Controller.

The rest is history: the handset hands over from the macro BSC to the UMA Controller using the standard GSM procedure and the call continues just as if it had handed over between two macro GSM cells.

Of course, I am simplifying some of the above but this is essentially what happens during the handover to WiFi/UMA procedure. Hand-out is much simpler and is essentially a handset driven process ... there is no configuration of the macro network required to support hand-out from UMA to GSM.

Anyone who is further interested can read all the gory details in TS43.318 and TS44.318 available free from the 3GPP website as UMA is a 3GPP Release 6 standard.

Again, hope this helps ...

IPobserver 12/5/2012 | 3:05:02 PM
re: AT&T's Foggy FMC Plans Hi kampar,

Do you know if there are any UMA services that offer hand-in to WiFi during the call?

IPobserver 12/5/2012 | 3:05:02 PM
re: AT&T's Foggy FMC Plans About handoverGǪ

Hand-out from the WiFi to Cellular is supposed to be easier/preferred than hand-in.

Forget the exact reason whyGǪ. I think it's partly because it's hard for the macro network (designed for a few 10s of thousands of sites) to know about potentially hundreds of thousands of UMA access points and, therefore, when exactly to release the call.

This is the case for femtocells, not sure about UMA. I'll look this up, or ask someone about it.

Another reason, I think, is that if the provider has differential pricing (it sounds like T-Mob has this), then the call needs to billed from where it was originated. I believe this is an issue for BT in the UK, which offers the cellular part via an MVNO.
IPobserver 12/5/2012 | 3:05:02 PM
re: AT&T's Foggy FMC Plans Thanks: It's good to hear about new wireless services that work well and people like.

Peter Heywood, a founder of Light Reading, signed up for BT's Fusion service when he retired and moved to the country (weak GSM signal). He says it works well enough.

I've never used UMA G partly because I like 'smartphones' with all the toys.
Page 1 / 2   >   >>
Sign In