LTE – The Voice Problem

Posted: March 23rd, 2011 | Author: Simo Isomaki | Filed under: Industry Insights | Tags: , , , , | 2 Comments »

A few months ago, I wrote a blog about whether LTE was breathing a new life into IMS. I received a lot of comments, and that post has been one of the most popular on the “The Dynamics of OSS”. Clearly, I touched a nerve there!

This time, I would like to look at how operators plan to solve “the voice problem” in LTE. Yes, voice! Given the revenue operators still make from voice services, one would have thought they would have made voice a priority. In fact, the focus so far has been almost entirely on faster data. However, the “voice problem” won’t go away.

The problem arises because 2G/3G voice is basically circuit switched, not packet transport. LTE, on the other hand, is all-IP, packet transport with no ‘circuit’ at all. This creates quite a problem for operators when they decide to transition voice from 2G/3G to LTE.

Currently, operators with LTE are considering three options to support voice:

  1. VoLTE (GSMA- and 3GPP-backed Voice over LTE, basically voice services in IMS core)
  2. VoLGA (Kineto Wireless-supported alternative Voice over LTE using Generic Access or UMA also known as unlicensed Mobile Access)
  3. CSFB (Circuit Switched FallBack, prior to Voice over LTE, the 3GPP-backed standard)

As it relates to voice calls, CSFB is principally not LTE at all.  The radio connection is moved to a circuit-switched 2G/3G radio connection (and LTE radio is not on). As a result, data sessions (web surfing, data streaming, etc.) would be cut unless the device has a dual radio mode (LTE and 2G/3G radio on at the same time), which apart from consuming more battery life, this would generally be a bad compromise.

VoLGA is a somewhat of a competing offering by Kineto among others. It proposes to bridge the gap between LTE and CSFB, but would require each handset and device to also support VoLGA. So far, the adoption of VoLGA appears to be limited.

This leaves VoLTE.  VoLTE seems to be the winner, as it is backed by all of the major network and device vendors. However, it mandates IMS back-end core and also forces all services in existence in 2G/3G voice (IN services, prepaid, roaming) to be re-implemented in IMS, which is clearly a big and potentially expensive challenge for operators. VoLTE would support the handover of radio to ensure voice call continuity in a single-radio mode (very good for sparing battery consumption), but some work is still needed in the standardisation for the handover process from 3G to LTE.

LTE could also naturally drive OTT (over the top) voice services (e.g. Google Voice, Skype, etc.) , but the challenge is that the handover to 2G/3G radio would cut the data sessions and disconnect voice. As LTE coverage will most likely not be 100% (at least for most European operators), this will most likely impact some users. That being said, it would only affect truly mobile customers. Nomadic customers, who are essentially stationary, would not have to face that problem and could use OTT services to the detriment of operator revenue.

However, it seems many operators have decided to tackle the challenge by first driving the rollout of LTE by using data only. Then, when coverage and hunger for more data is handled—and handled well—the assumption is that voice can more easily be put onto the network. I’ve heard rumours that larger international groups are building a central IMS core to be shared between their national operating companies (the LTE radio part would remain separate). This would help to drive the business case for IMS, as IMS costs are split between the operating entities. However, sharing IMS rollout costs perhaps will not be available to smaller, independent operators then.

With CDMA, the world is ’simpler’ as the 3GPP2 standard for handover procedure from LTE to CDMA is not available, and at least Verizon is not even looking into that.

Needless to say, all of this will have an impact on back-end systems, like charging and policy control, as well as service creation and activation, which will require special attention—especially in the complex world of hybrid LTE+2G/3G networks, where services still need to be created and charged for so that they seamlessly work during handovers.