By Dean Bubley, Disruptive Analysis

Using 3G networks for VoIP is inevitable, and will be more important than VoWLAN.

Future generations of wireless technology – LTE, UMB, WiMAX – are all-IP, so unless mobile operators continue to run separate voice networks in parallel, they will have to transition to VoIP eventually. But these radio technologies are 3-5 years away from mainstream deployment, and in the interim, the situation is much less clear. Today’s 3G networks can already support VoIP, albeit often without the quality, efficiency and reliability of circuit-switched telephony – and without the awareness and hype garnered by WiFi VoIP.

While VoWLAN has many advocates, and has compelling advantages in environments like enterprise offices, it has some severe limitations. Most obviously, WLAN coverage is patchy and access is often locked-down by individuals or corporations. In addition, software complexity on the handset for managing WLAN is significant. Operator-controlled architectures like UMA have struggled with myriad implementation difficulties, and have had only mediocre uptake. While these problems are slowly being solved, VoWLAN remains more important qualitatively  than quantitatively. While it is encouraging innovation and new business models, it is having very little impact on the wider mobile marketplace.

Some independent VoIP players are already extending their mobile VoIP software on smartphones to exploit 3G as well as WiFi. At the same time, there is an increasing trend of carriers marketing 3G modems for PCs – not just for mobile computing, but also to compete with DSL/cable broadband offerings. Laptop users expect to be able to use their normal broadband applications, including voice-based ones like Skype or Facebook plug-ins.

More conservative mobile operators generally wait for standardised, fully-optimised technologies to become available before large-scale deployment. For CDMA carriers, telco-grade VoIP should be available relatively soon, as EV-DO Rev A has long had packet voice ‘designed-in’ as an expected primary usage case. Most of the larger CDMA operators also have the luxury of very tight control of their handset architecture, enabling homogenous software and chipsets to be rolled out quite quickly. There is also latent demand for VoIPo3G-based push-to-talk, especially in the US, which should help ease the path towards full CDMA IP telephony at a later stage.

Conversely, Disruptive Analysis believes that 3GPP operators with UMTS networks face some serious challenges. The current newest variant, HSPA, was not originally designed with VoIP in mind as a core usage case. Although it is voice-capable, it is not voice-optimised, unless all its numerous add-on options are actually implemented –not the case in most existing deployments. The next, more VoIP-centric generations HSPA+ and LTE are still several years away, and dependent upon new spectrum allocations.

But operators are also unlikely to want to risk rolling out a complex new radio network (LTE) and transitioning their core telephony service to VoIP at the same time. Furthermore, while the standards bodies have created “IMS multimedia telephony” as an enhanced service, they have neglected to develop ‘plain vanilla VoIP’ as a mobile-centric application so far. To gain some pre-LTE experience of packet voice, carriers will therefore be forced towards pre-standard or proprietary approaches, with attendant risks such as limited handset range or poor interoperability.

Importantly, HSPA networks often tend have a growing proportion of users with ‘open’ smartphones, as well as growing numbers of 3G-enabled laptop subscribers. High levels of competition in many HSPA markets are driving down data charges towards flatrate. Together with technical limitations and regulatory oversight, this is likely to limit carriers’ attempts to block or differentially charge for ‘off-portal’ VoIP traffic.

Again, the situation in the CDMA world is rather different, as there are far fewer fully-open smartphones, and handsets are often heavily operator-customised. Nevertheless, in some CDMA markets like the US there is also a regulatory and market push towards device and application openness, as evidenced by the FCC’s rules for the 700MHz auction, Sprint’s WiMAX plans and Verizon Wireless recent u-turn on supporting 3rd-party devices and applications.

Disruptive Analysis believes there is an argument for HSPA operators to partner with the independent software and Internet VoIP players. They need a way to get their own foothold in VoIPo3G; they cannot afford to sit back and allow their competitors to exploit a 3-4 year window of opportunity unchallenged. They also need a way to gain experience of VoIP before LTE rollout. They may also be able to exploit new IP-based service opportunities, such as mobile voice mashups.

In return, independent VoIP companies need operators’ scale in device distribution, service management and marketing: many mobile users shy away from downloading 3rd-party applications to their phones and self-configuring them. Partnership models of various types represent a potential win-win, pre-loading and pre-configuring VoIP applications so that they are usable when the customer leaves the shop. This can drive uptake (or upgrade) of data plans and enable a variety of ‘smart pipe’ business models. The recent 3 / Skype deal is an early example of this general philosophy, although in its current version it does not involve VoIPo3G yet.

For CDMA operators, the VoIPo3G rationale and roadmap is somewhat different. They are on track for VoIP readiness anyway, and are likely to be aggressive with their own deployments, starting with large-scale 3G-based push-to-talk. Some like KDDI and Verizon are also aiming for all-IP core networks and FMC VoIP applications, spanning both their mobile operations and fixed business units.

Overall, the tracks of CDMA and UMTS worlds - and of carriers and ‘over the top’ independent providers - are actually converging. End-to-end VoIP is becoming mainstream in the fixed world, and is creeping into wireless though VoWLAN. Over the next five years VoIP will start its migration to wide area mobility. It is likely to be at least 2015 before a majority of cellular voice minutes are based on VoIPo3G or VoIPo4G, but the next few years will set the scene for how the eventual transition will occur. Most of the 255m user base of VoIPo3G in 2012 will still use circuit voice as well – but the ways in which they purchase and use early mobile VoIP will define the shape of future value chains and mobile operator strategies.


This article is based on Disruptive Analysis’ new report on Mobile VoIP. For more details, see www.disruptive-analysis.com .  Dean Bubley is the author of the report, and also writes at disruptivewireless.blogspot.com