Operators Raise Voice Services on LTE

With the latest launches of Long Term Evolution (LTE) smartphones, operators have implemented another way to deliver LTE voice services in the network, called circuit-switched fallback (CS fallback).

To date, only CDMA operators like Verizon Wireless and MetroPCS Inc. (NYSE: PCS) have supported voice calls on LTE smartphones by having more than one radio active in the device -- one for the packet-based LTE data service and another for circuit-switched CDMA voice service.

Now, GSM operators like AT&T Inc. (NYSE: T), Rogers Communications Inc. (Toronto: RCI) and soon Telia Company can also support voice on their LTE smartphones. The LTE devices will have multiple radios to support LTE, UMTS and even GSM, but with CS fallback only one radio can be active at any time. So, the device is forced off the LTE network and onto the 2G or 3G network for voice calls.

AT&T launched LTE devices earlier this month, while Rogers activated its first LTE devices at the end of October. Both operators have launched CS fallback, according to Gabriel Brown, senior analyst at Heavy Reading. (See AT&T Readies Its First LTE Phones .)

In addition, TeliaSonera announced at the end of last week its first LTE tablet, the Samsung Electronics Co. Ltd. (Korea: SEC) Galaxy Tab 8.9 LTE, which will go on sale before Christmas, as well as a Samsung LTE smartphone that will launch early next year and support CS fallback for voice services.

"We will not sell the smartphones until the CS fallback is done. And this isn't an issue," said Tommy Ljunggren, VP of system development for Nordics and Baltics at TeliaSonera. "That is what Europe has been waiting for -- it's crucial to have CS fallback."

According to Heavy Reading's Brown, a majority of operators are launching CS fallback because it's a "time-to-market issue" before voice over LTE (VoLTE) will be commercially available, which Brown says will be in the first quarter of 2013.

Until then, operators can opt for CS fallback, which has been viewed as an interim solution.

"The question becomes, Is there such a thing as an interim solution for CS fallback or will it continue to run in parallel with VoLTE?" says Brown.

TeliaSonera's Ljunggren says operators will need both CS fallback and VoLTE. "You can't live with only one of them," he said. This is because CS fallback will be needed to support users as they roam outside of 4G coverage areas. "It's the most natural way to go forward right now."

For now, though, there are more practical issues about CS fallback for operators to contend with, such as reducing voice call and data session setup times.

"Roughly speaking, the first CS fallback implementations add a second or two of extra call setup time," says Brown. "But there are enhancements coming next year in the network and terminal that will get that down to about 500 milliseconds [of additional call setup time]."

There is also another optimization technique called network-controlled packet-switched handover, which will work in parallel with CS fallback, that will reduce the data session setup times down to a few hundred milliseconds, according to Brown. This will be available next year.

— Michelle Donegan, European Editor, Light Reading Mobile

Page 1 / 2   >   >>
Kevin Mitchell 12/5/2012 | 4:47:52 PM
re: Operators Raise Voice Services on LTE

Actually, both the CDMA-LTE and GSM-LTE CSFB are very similar and both have 2 radios (or more) in the device.

In the CDMA approach, both radios are active at 1 time: LTE for data and CDMA 1xrtt for voice.

In CSFB (only a standard for the GSM 3GPP world), only a single radio is active at a time: primarily LTE. However, when a call is made or received, the phone is kicked off LTE and attaches to the GSM network for the duration of the call. Its call related, not data. I'm not sure how the dual radio device knows to attach to UMTS/HSPA for data when out of range.

It seems to me that this brings the CDMA "charm" of single sessions to the GSM world: only one voice or data session (not both) is active at a given time. Sounds like a STEP BACKWARD to me. Is this wrong?

CSFB is NOT needed for handover with call continuity. That's a bit different and requires coordination between IMS and MSCs. Single radio voice call continuity is something being refined in 3GPP now. That requires new IMS elements and MSC upgrades.

Gabriel Brown 12/5/2012 | 4:47:51 PM
re: Operators Raise Voice Services on LTE

It does in some ways seem like a step backwards, but the data session does/can restart when the device falls back to UMTS. Potentially CSFB could be combined with network-controlled PS handover to optimize this.

It'll be interesting to see how much of an interim solution CSFB is. Once it's installed and devices start shipping it'll be in service for a long time. I wonder how and when VoLTE will take-over as the primary voice solution?

Agree that "CSFB is NOT needed for handover with call continuity". That's two seperate issues.

paolo.franzoi 12/5/2012 | 4:47:50 PM
re: Operators Raise Voice Services on LTE


You know its funny.  My UMA phone worked just fine several years ago.  Odd that the technology could not be re-used.



rduncan 12/5/2012 | 4:47:50 PM
re: Operators Raise Voice Services on LTE

Clarification to article - Both CDMA-LTE and CSFB require two radios:

CDMA-LTE uses (1) LTE and (2) CDMA radios

CS FallBack uses (1) LTE and (2) GSM/UMTS radios

In both cases, when LTE is not available, the handset falls back to the second radio for ubiquitous radio coverage (that LTE will lack for some time).

  Robert (www.imcellular.org)

Dan Warren 12/5/2012 | 4:47:48 PM
re: Operators Raise Voice Services on LTE

Ok folks, lets get this straight.  Both CSFB and the 'dual radio' solution that VzW and the like have deployed require two radios - LTE plus whatever the legacy happens to be.  However, they do not operate in the same way.

CSFB first - I am on LTE radio and someone calls me.  The LTE network tells my phone that a call is being made, but because my device has no voice capability on the LTE radio access, the network and device between them force my phone back on to a 'CS capable' bearer - so either WCDMA (in one of its variants) or GSM.

CDMA/LTE dual mode devices can be attached to LTE, but have the CDMA-2000 based radio connection active as well.  They are essential attached to the network twice.  If a call is made to me, then the LTE radio is not detached - the call just routes to my already active CDMA-2000 connection with the LTE radio being unaffected.

So as Gabriel says, similar but different.  The difference is not when the device is out of LTE coverage (when both phones act as a non-LTE device does today), but when the phone is in LTE coverage.

Both solutions require specific support in the device, and as a result any network implementing these solutions will have to support them for as long as devices supporting the techniques exist in the network.  Call set up in the dual-mode CDMA/LTE version is the same as existing call setup (since the network just thinks it is establishing a call to a CDMA device), whereas calls etup is longer by a small number of seconds (which can feel like a long time) with CSFB.

I am also inclined to agree with Gabriel that the longevity of an 'interim solution' is much more than 'until the day when VoLTE starts'.  Devices are in the hands of consumers, and consumers don't necessarily all change phones that often.  So swapping out an interim solution is either a very long process or requires proactive, and potentially costly action on the part of the Operator to go and replace phones that are in the hands of people that wouldn't otherwise be swapping.  Either way, it remains my view that if an operator can swap straight into VoLTE as the long term solution, without an interim solution it is probably a 'cleaner' evolution in that it doesn't require something to be introduced in the network and devices that you already know you want to get rid of.

The issue is simply one of timing, but if enough focus were placed on VoLTE by operators, handset manufacturers and network equipment manufacturers, the need for CSFB could have been minimised.

Gabriel Brown 12/5/2012 | 4:47:48 PM
re: Operators Raise Voice Services on LTE

My sense is that CSFB is out in the wild and, therefore, will proliferate. I think most operators thought they'd have to do it anyway, and now the optimizations are coming through it seems a decent option.

For now, I don't think this is causing people to delay work on VoLTE -- in fact several operators who already have LTE networks I've spoken with recently are comfortable doing both. There are so many moving parts to VoLTE that it will take time (goes right back to the chipset) to fully implement. If you want to start selling devices, you have to go with CSFB (or SVLTE in CDMA).

But VoLTE is still the target for primary voice service for virtually everyone.

Like Dan says, it's a question of timing. I would add the "how" to that as well -- best-practice processes for implemenation, interconnects, roaming, and so forth are critical. Dissemination of this kind of information through the eco-system is vital and will impact the schedule as much as pure technology.

A lot of people (especially outside the Tier 1s) are looking to the GSMA to do the leg work on this, which to its credit, it seems to stepping up to. Vendors too are important in this best-practice information sharing.

Might as well insert a plug for a webinar we're hosting on December 7 on VoLTE. It will cover lots of these issues with input from four vendor speakers all close to VoLTE evalutions & implementations in various operators.


It's free to attend and participate in this session.


Kevin Mitchell 12/5/2012 | 4:47:47 PM
re: Operators Raise Voice Services on LTE

Dan, what about this issue on CSFB?

It seems to me that this brings the CDMA "charm" of single sessions to the GSM world: only one voice or data session (not both) is active at a given time. Sounds like a STEP BACKWARD to me. Is this wrong?

Can you also elaborate on the network readiness? Yes, it's a standard, but what about vendors, networks

I don't think many networks are ready today for CSFB. It requires functionality in MSCs and MMEs that are not necessarily there in today's code. Are the MSC vendors ready with the code updates they need? How extensive an effort is this for service providers?

What about handsets? When do they become commercially available>

Michelle Donegan 12/5/2012 | 4:47:47 PM
re: Operators Raise Voice Services on LTE

Thanks, everyone who pitched in with explanations. I think I get the difference now between CS fallback and LTE voice on CDMA (SVLTE): both use multiple radios in the handsets. With CS fallback only one radio is on at a time. With the CDMA approach, both radios are on all the time.

That about right?



Kevin Mitchell 12/5/2012 | 4:47:46 PM
re: Operators Raise Voice Services on LTE

Some quick thoughts on this.

Why UMA/VoLGA lost to VoLTE/IMS:

<li>offered only service parity and not multimedia communications based on SIP</li>
<li>small ecosystem of supporting vendors</li>
<li>requires operating and investing in expensive/aging legacy MSC core</li>
<li>It had a vulgar name</li>

Why UMA/VoLGA lost as an interim solution to CSFB:

<li>VoLTE/IMS vendors pushed CSFB</li>
<li>Requires IPsec clients on handsets which has proven difficult with negative impacts to battery life and time to marke</li>
rduncan 12/5/2012 | 4:47:46 PM
re: Operators Raise Voice Services on LTE

&gt; My UMA phone worked just fine several years ago.
&gt; Odd that the technology could not be re-used.

UMA was a strong candidate to support Voice where LTE was not available, and I liked that alternative, too. UMA is Proven technology, available wherever there was 3G data (so the fallback option had the same fallback as Circuit-Switched Fallback, as I dount that there are many 3G locations that have only Circuit and not Data!). T-Mobile led a group of Operators to support it. It lost to Circuit-Switched fallback.

Anyone know why UMA lost to CSFB? Was it primarily the incremental cost of implementing UMA?

&nbsp; Robert

Page 1 / 2   >   >>
Sign In