Mind the gap

By Pamela Weaver

Data might be the driving force behind LTE’s gathering pace, but there’s no escaping the fact that voice continues to be the killer application for mobile operators. “Smart” they may be, but drop the “phone” element from the current crop of devices and you’re dropping 64 per cent of your revenues—according to Ovum research, this is the percentage of MNO turnover that voice will generate by 2015.

LTE standards were developed from the getgo on the understanding that circuit-switched voice was not going to be part of the overall picture. This means that mobile operators are faced with the task of bridging the gap between their network evolution strategies and current 2/3G realities.

It’s partly their own fault: back in 2002, the 3GPP identified IP Multimedia Subsystem (IMS) as the specification for packet voice. Since then most of the equipment vendors have developed IMS implementations, and they’ve had them for some time. But expensive- to-implement IMS languished on the sidelines while operators focused on LTE projects, dropping the voice ball in favour of tackling the seemingly more immediate threat from a mobile data explosion and the slew of over-the-top (OTT) players that followed in its wake. It was always just assumed that issues such as global scale, interoperability, interconnect and roaming would somehow have been dealt with by the time LTE was ready. Until recently, they weren’t.

The GSMA’s VoLTE initiative was launched in February 2010 and has subsequently been involved in defining the protocols and standards required to make voice over LTE work harmoniously across devices and networks. IMS is a central component of making it work. “The problem we were always going to have with IMS was justifying it as a technology from a business case perspective,” says Dan Warren, senior director of technology at the GSMA.

Now that the VoLTE initiative has brought IMS back to centre stage (VoLTE is essentially voice services in the IMS core), perceptions that IMS is too complex or pricey have given way to expediency. “Voice on its own justifies the expense,” says Warren. “It’s a hundreds of billions of dollars market that needs to evolve to something new. IMS is the best game in town for the carriers to do that with. That alone justifies IMS as a future-facing technology.”

Given that we’re some years away from anything like blanket LTE coverage, roaming and parallel network maintenance will be the name of the game. Many telcos will use LTE for data and legacy networks for voice on a circuit-switched fallback (CSFB)/single radio voice call continuity (SRVCC) basis. CSFB has attracted its share of criticism, with Steve Shaw, vice president for corporate marketing at infrastructure player Kineto Wireless calling it a “truly horrific solution” that gives early LTE device adopters a “whole reason not to take a call—and that’s your primary revenue lifeblood.”

Be that as it may, for most operators right now, voice is a 2/3G play and will be for the foreseeable future. Ovum analyst Steven Hartley notes that the likes of US carrier MetroPCS and Japanese incumbent NTT DoCoMo are both following a CRFB transition path in the early stages of their LTE deployments. “LTE is used for data, legacy networks for voice,” he says. “It involves maintaining parallel networks, but you’re going to need to do that anyway until LTE coverage and device penetration is sufficient to switch them off.”

Tommy Ljunggren, SVP of system development and mobility services at LTE pioneer TeliaSonera, has said that his firm’s 4G launch was made easier by the fact that TeliaSonera “focused on data only and wasn’t complicated by voice and voice integration with legacy systems and so on.”

Meanwhile, Ericsson’s head of Mobile Telephony Evolution, Eric Ericsson, says that the most common evolutionary path he sees operators choosing involves CSFB in some way, not least because it’s the industry choice for supporting inbound LTE roamers who don’t have IMS services or roaming capabilities.

“We’re seeing operators introducing LTE with enabled handsets and then opting for CSFB in the initial stages while coverage is spotty,” he says. The natural progression for most of these operators will be to upgrade to VoLTE one-to-two years later, using SRVCC to manage handover between LTE and the circuit switched network. It’s an interim measure, yes, but as Ericsson points out, carriers have already invested heavily in their 3G networks. Given that a primary driver of LTE implementation is cost rationalisation in core access, throwing the 3G baby out with the bathwater doesn’t currently make any sense.

Ericsson says that many of the implementations his company is involved in re-use the existing legacy environment of the MSC, “breaking out to the legacy world so our existing mobile media gateways remain in the network, even for VoLTE.” Ultimately, when LTE has broader reach, fallback will be onto the HSPA or 1xEVDO network— “that’s the all-IP scenario,” he says. “The availability of spectrum steers the technologies.”

For Ericsson, it’s all about evolving the mobile core while maintaining a connection to the legacy, rather than replacing it completely. Some might not like it—the GSMA’s Warren, for example, says that interim steps between CS and IMS will be difficult to switch off in the long term—but with Ericsson predicting that “there will be circuit-switched users on the mobile side beyond 2020,” it’s clear that there is going to have to be some degree of pragmatism in operators’ LTE strategies going forward.

Four myths about LTE
Stephen Hayes, principal engineer at Ericsson and chair of the 3GPP’s systems group has said there are four myths surrounding LTE and its capacity to carry voice. According to Hayes, these myths—and the realities—are:
• LTE is data only: Support for voice was one of the key considerations in designing LTE. The solution for
LTE is IMS VoIP and it is fully specified.
• SMS isn’t supported over LTE: LTE and EPS (evolved packet system) will support a rich variety of
messaging applications—including SMS. The solution is twofold, covering both the full IMS case and a
transition solution for those networks that do not support IMS.
• IMS isn’t ready for prime time: IMS was first developed as part of 3GPP Release 5 in2002. It is based on
IETF protocols, such as SIP and SDP, that are very mature. These technologies have been embraced by
the industry as the signalling mechanism for multimedia applications.
• LTE doesn’t support emergency calls: VoIP support for emergency calls (including location) are included
in Release 9.
A transition solution—fallback to 3G/2G has existed since IMS was introduced in Rel5.

While CSFB will be the way forward for many operators transitioning to LTE, carriers that are pursuing a more aggressive path to the future, such as Verizon, are leaping with both feet into an all-IP future. Rather than looking at 100 per cent backwards compatibility, the telco is evolving straight into offering services that will compete with other OTT communications services. Having demonstrated one of the world’s first VoLTE calls at this year’s Mobile World Congress on Samsung’s Revolution handset, the company is gearing up for the commercial launch of VoLTE-using-IMS in 2012. AT&T will follow suit by 2013; T-Mobile USA, according to Warren, will be sticking with HSPA+ for quite some time “because their last RAN investment gives them that evolutionary path via software but doesn’t allow for an LTE upgrade without a hardware investment.”

Verizon’s more direct migration path can, in some ways, be attributed to its existing CDMA network, which isn’t easily compatible with LTE. But it’s also likely to be part of a broader industry anxiety about OTT players and the damage that some observers believe they’ll be able to inflict on operators who spend too long hanging around on CSFB. Circuit-switched fall back sceptics like Kineto’s Steve Shaw say that, by implementing it, operators are opening the door to providers like Skype who can offer full video and audio calling over LTE’s super-fast, super-efficient technology. Shaw maintains that subscribers will simply stay on LTE, using OTT services for voice, getting annoyed with CSFB-induced delays and gradually eschewing carrier services for calls.

For the GSMA’s Warren, however, it’s a glass-half-full/half-empty scenario. While he believes that it is unrealistic to even try to pretend that OTT voice is going to go away, there are plenty of drawbacks to voice offered on an OTT basis, not least quality of service. As Acme Packet director of solutions marketing, Kevin Mitchell says: “Left alone, IP networks treat every packet the same.”

Policies need to be put in place to ensure QoS for latency-sensitive services like voice and video, giving them priority and the bandwidth path they need. “Even though LTE gives us blazing-fast bandwidth, there are different constraints and choke points sill in place,” says Mitchell. “It doesn’t mean resources can’t be overloaded in the core network.”

Engineering around that and implementing session border controllers can handle those issues, ensuring the QoS paths are there in a way that OTT players simply can’t.

And if Kineto’s Shaw thinks users won’t be happy having their video streaming interrupted to take a call, Warren believes they’re even less likely to tolerate a bad voice service. “Unless it’s an OTT service that fully integrates and works out of the box, there are still a lot of people out there who want high-quality voice provided by an operator,” he says. One of the best things about the Verizon demo, according to Warren, was the use of AML wideband and high-definition voice, something he believes end users will see as a real benefit going forward. Kevin Mitchell agrees, saying that IMS architecture is carrier-grade, not OTT or best effort.

“There’s a view out there that voice has become the domain of OTT, some kind of arbitrage or dial-around that doesn’t involve the service provider in the revenue stream, but IMS is VoIP,” he says. “Service providers using IMS have control over the access network, policy control mechanisms…they’re deploying different technologies throughout the access and core networks to make sure its carrier grade.”

According to Mitchell, OTTs can bring a lot of innovation to an all-IP environment, but at basic levels of QoS, compliance and emergency services offerings, they’re not in anywhere like the same category as traditional operators. As Eric Ericsson puts it, OTT services are “at the telephony and service layer, but aren’t specified in using the lower layers for optimised media handling, policy control and radio capabilities. That’s why VoLTE, in terms of robustness, will work differently and better than a service running OTT.”

While Mitchell sees a future in a service federation model—where OTT players and carriers such as Skype and 3/Verizon/KDDI offer hints of how the two sides might cooperate at the network as well as the business level—Warren argues that there’s plenty of room for operators to go it alone on the VoLTE front. “Companies aren’t going to invest in the IMS core just for voice,” he says. With operators evolving their service offerings in order to compete, it’s only a matter of time before people are persuaded of the value of voice calls plus the entire IMS-enabled infrastructure built around it, he argues.

“Give subscribers the capability to do instant messaging in a way similar to SMS, give them presence information and data relating to what the person they’re calling is able to do with their phone—like a classic rich content suit (RCS) model of a contact list—promote interaction between users and their contact lists in ways they might not otherwise have done,” says Warren. “Ultimately that drives up call volumes.” And, presumably, revenues.

For Ovum analyst Steven Hartley, while voice is becoming commoditised and isn’t generating revenues in line with volumes, neither is data. The key strategic objective as voice moves on to LTE, he says, will be “managing costs effectively to ensure margins are protected, even if revenues are under pressure.”

Ultimately, says Hartley, operators will charge access fees that include bundles of minutes, text, data and whatever else they want to throw in. What proportion goes to voice or any other element is actually irrelevant, as long as the whole is making a profit…”

In many ways, VoLTE represents a real opportunity for carriers to transition into the smart pipe stage by differentiating themselves on voice and services related to it. As Dan Warren puts it, operators are definitely chasing more bandwidth. In terms of how or what they do with it, there aren’t really any right or wrong answers, just “answers based on each individual market and their spectrum availability and their map for technical evolution.”

For operators calling time on carriers’ relevance in the face of CSFB vs OTT, it’s worth taking a couple of steps back. Yes, there are a lot of players moving very quickly but devices are limited and, as Warren points out, they’re essentially de-bugging LTE on the fly. If VoLTE isn’t 100 per cent there, it’s only fair to point out that neither are the LTE deployments or the handsets that will take advantage of them. Sounds like a good time to bridge the IMS gap.