Your browser is not supported. Please update it.

4 December 2007

Cellcos overcome SIP fears and VoIP over 3G set to dominate by 2012

Mobile operators, especially in Europe, have been conservative about moving too rapidly towards full IMS (IP Multimedia Subsystem) but increasingly they are seeing that they can take a first step by implementing VoIP, as a stage on the road to the full IP vision. This mentality, which has already taken strong hold among the US majors, will drive increasing commitment to SIP as the main basis of converged and IP services, rather than the main alternative, dual-mode Wi-Fi, and will create a boom in use of VoIP over 3G.

According to new figures from Disruptive Analysis, this boom will see VoIP over 3G users rising in number from almost none this year to over 250m by the end of 2012, eclipsing voice over Wi-Fi and leading to rising investment in carrier class VoIP systems, as cellcos seek to keep control of VoIP usage, rather than losing the customer to unlicensed Wi-Fi, or encouraging use of open VoIP clients like Skype. To date, European operators in particular have been wary of opening their networks too freely to applications like Skype, although some are working to bring such apps under their own control with versions optimized for their own handsets, an approach favored by 3.

Start-ups like Tatara Systems and Sonus have been betting on the acceleration of uptake of SIP-based services, well in advance of wholesale IMS deployments, and particularly that VoIP in the 3G core will be an initial step in this direction. They seem to be seeing their bet pay off, with larger players like Alcatel-Lucent joining the market, and operators like AT&T talking up their plans.

Traditionally, VoIP has used far more bandwidth than circuit switching, especially on the highly efficient 3G voice systems, but the latest network upgrades and the introduction of IMS make cellular VoIP as efficient as circuit switched in bandwidth terms. And there are advantages such as greater utilization of the spectrum, since voice is delivered as packets, rather than requiring a dedicated channel; lower operational costs; and better integration with other IP features, supporting premium add-ons such as instant messaging. Dean Bubley, founder of Disruptive Analysis, also points to the ability to reduce operating expenses by combining fixed and mobile core networks, and to use VoIP as the springboard to launch other new services like push to talk and voice-integrated mash-ups. VoIP over 3G is also a strong fit with the rising interest in extending the RAN, and improving indoor penetration and homezone schemes, by using femtocells within the home.

Operators will inevitably transition to IP networks in a few years’ time ‘ mainly LTE or WiMAX ‘ but in the mean time, there is a pressure to act earlier, particularly in the laptop sector. Many carriers are now marketing 3G laptop modems as a more flexible alternative to a fixed home broadband line, and like that line, this option needs to come packaged with carrier branded VoIP, to increase the value proposition to the user without encouraging the strain on the network of high levels of Skype usage. By 2012, the report claims most VoIP over 3G users will be on mobile carriers’ own standards-based VoIP systems, although a large minority ‘ about 60m ‘ will prefer independent offerings, many actually operated in partnership with cellcos.

“3G networks are increasingly capable of supporting VoIP, for both traditional mobile operators and independent internet-based VoIP challengers. But while CDMA operators will benefit from VoIP being ‘designed-in’ to their newest networks, 3GPP/HSPA operators will have to wait for several years – a window of opportunity which will be exploited by the ‘over the top’ players. Rather than competing head-on, partnership models have the potential to create win-win propositions,” said Bubley.

While cellcos can no longer afford to ignore mobile VoIP, with a rising number of start-ups coming to market with low cost offerings based on hotspots, it is clearly a double-edge sword, squeezing the occasions in which consumers will have to pay the cellular premium. As with data hotspot access, mobile carriers know it is better to control Wi-Fi and use it to attract and retain cellular customers, than to see their revenues eroded by a new challenger, but they face a potential nightmare of expensive 3G networks being underused for the application they do best, voice, and the necessity of creating a second Wi-Fi or WiMAX system to support VoIP. This is driving the operators towards running VoIP over the cellular networks, both to optimize usage of those dearly bought links and to make it unnecessary for consumers to move off the cellular system. Vodafone Germany was among the first to harness the spectral efficiency of its 3G network to deliver voice at low cost, flat rate tariffs that could take on the low cost voice over Wi-Fi providers; and Verizon Wireless has been one of the most aggressive about running VoIP over its emerging CDMA EV-DO Rev A system.

In other research, conducted in 2006, Analysys found that revenues from cellular VoIP would overtake those from fixed line voice by 2012, and this approach would sideline VoIP over Wi-Fi and other wireless systems (although to be fair, by that date we will be close to the all-IP 4G network and it will be hard to distinguish between ‘cellular’ and ‘broadband wireless’ technologies any longer). Analysys predicted that cellular VoIP services would generate revenues of $18.6bn in the US and $7.3bn in Western Europe by 2012, compared with fixed VoIP revenues of $11.9bn and $6.9bn respectively. The upgrade to CDMA EV-DO Revs A, B and potentially beyond, and of W-CDMA to LTE, both of which will support full IP and VoIP, will make the business case for cellcos to move away from circuit switched voice, as will improvements in the performance of cellular VoIP, and the opportunity to position it as a premium service via value added features like instant messaging. There will also be large numbers of operators – either new to the wireless field, or cellcos that have adopted multiple technologies, deploying VoIP over WiMAX or Wi-Fi.

“The capacity, cost per megabyte and quality of service of existing 3G cellular technologies – including EV-DO Revision 0 and HSPA – are not yet adequate to support a significant move to wireless VoIP services,” said analyst Mark Heath. “However, EV-DO Revision A and 3G LTE will respectively create the cost benefits and new service opportunities that trigger the migration to mass market cellular VoIP.”

T-Mobile may be a flag waver for multinetwork services, but Verizon Wireless has been the most aggressive in support of cellular VoIP plans, partly because its chosen 3G network, EV-DO, will enable VoIP ahead of Cingular’s UMTS. CTO Dick Lynch, like his counterpart at fellow CDMA carrier Sprint Nextel, Barry West, thinks upgrading to Rev A will achieve a 35-40% capacity increase as well as supporting VoIP, making the proposition far more appealing. Sprint is holding fire on VoIP on its CDMA network, partly because it is also building out WiMAX, but Verizon plans to use VoIP over cellular as early as next year for selected premium offerings such as enhanced push to talk. Another vital advantage will be to increase its showing in the large enterprises, which increasingly want all their communications, including mobile voice, integrated under one IP umbrella. Rev A will also enable some traffic to be offloaded from the current CDMA network, and will support simultaneous voice and data calls – including applications like conferencing – and hand-off to Wi-Fi.

“When we get Rev A, we have a pure IP platform that can do VoIP at efficiency levels at least as good as 1X,” Lynch said in an interview. This will be a step towards full all-IP convergence based on IMS, an area where Verizon Wireless has taken a lead in bringing together major vendors to create an enhanced IMS design (AIMS) tailored for the needs of a CDMA-based mobile carrier.

W-CDMA operators such as AT&T have had less incentive to deploy VoIP at an early stage, unless they needed to integrate voice calls with a fixed network or hotspots, but are now moving quite rapidly towards SIP and VoIP. VoIP is possible over W-CDMA, but the network would lose 80% of its capacity, but HSxPA reduces that to 30%. Most W-CDMA operators are waiting until they implement 3GPP Release 6, which features an interference cancellation scheme (GRAKE) which should give VoIP over cellular the same voice capacity as current W-CDMA. The next generation system, LTE or 3.9G, is promising to improve that to deliver VoIP at double the efficiency of conventional systems.

Until these major upgrades, the user demand to make VoIP calls anywhere, rather than just when in reach of hotspots or fixed lines, will be met by some stopgap VoIP over cellular options. These include specific applications that use a variant of the technology, notably push to talk, which uses the GPRS data channel on a GSM network to make a voice call. Also, internet-enabled cellphones can download VoIP clients such as Skype and also use the GPRS or EV-DO data channel when out of reach of a hotspot. However, many operators have not yet moved sufficiently far beyond their walled gardens to permit this practise in their contracts, although as with Wi-Fi itself, they are starting to realize they need to accept the reality of Skype, however painful to their original business plans – either by partnering with the company, as 3 and others have done, or by launching their own VoIP equivalents with strong pricing and superior added value features. The cellcos will, however regretfully, embrace VoIP and deliver it effectively, snatching a large part of the market for themselves, but as with internet data, they will face a wide range of competitors including all the wireline convergence players. Over the next two years they will deploy the technologies that will allow them to survive even in a world where they will never again have an entirely exclusive, tightly controlled technology, as they did with mobile circuit voice.