VoIP over 4G (Vo4G): Opportunities, Challenges and Deployment Trends

The communications industry has been working to realize fully converged networks for nearly a decade. The benefits are obvious: instead of running separate networks for each application, a single network capable of running all the applications provides increased efficiency and flexibility. Furthermore, new applications can be envisioned, created and deployed without the need for modifications to the network or worse, the deployment of entirely new networks.

In the wired domain, this transition is nearly complete. Worldwide, operators are deploying packet networks to deliver the big three applications: voice, Internet and video. The wireless networks however, still operate much like the wired networks of a decade ago because the technological challenge is so much greater in the wireless domain, especially for voice communication.

Voice communication is complicated. From a network implementation standpoint, the only simple aspect of voice is that it is low bandwidth. Every other aspect is difficult. Voice requires low delay, is bidirectional, is mostly continuous, and there are a host of features which voice users expect, and in some cases demand. For this reason, the majority of wireless networks today continue to deliver voice in the traditional manner, over a completely separate TDM network. The implementation of packetized voice, or Voice over IP (VoIP) on wireline networks, although complicated, was simpler than it has been for wireless, because available bandwidth is so much higher and delay is lower on wireline networks. Overcoming these obstacles for wireless has required a revolution in technology that has only begun to be deployed over the last year or so.

WiMAX and LTE promise to be the first wireless access standards fully capable of supporting VoIP. These two very similar technologies are the first standardized carrierclass technologies to have incorporated low delay along with sufficient Quality of Service (QoS) capabilities and bandwidth. In both, there are a rich and flexible set of features offered at the medium access control (MAC) layer for the construction and transmission of MAC protocol data units (MPDUs) to support multiple VoIP streams, but voice quality requires an end-to-end solution. Although the key parameters are available in these new technologies, delivering a voice access system on an end-to-end basis remains a technically challenging task.

While much care was taken in the WiMAX specification process to ensure that VoIP can be supported gracefully, WiMAX does not today specifically address how QoS is applied for voice traffic. It is up to the equipment vendor to decide how to classify packets as VoIP packets and how to apply QoS to those packets. Furthermore, because wireless VoIP is new in general, there are really no well-defined "best practices" for network architecture and provisioning. For these reasons, vendor selection is incredibly important, and VoIP performance on WiMAX has varied widely.

LTE will begin its deployments as the next architecture development in an industry that is orders of magnitude larger than the WiMAX industry. Whereas the 802.16e-2005 standard was mainly an air interface standard, LTE benefits from years of work on the core network. Furthermore, because it is an extension to an enormously successful 2G and 3G networks architecture, it will be possible to leverage the existing networks in the short term to ensure a successful transition. For example, during the transition when LTE is first deployed and coverage is spotty, a GSM/HSDPA/LTE multi-mode handset could use LTE only for hot-spot coverage in high demand areas, and fall back to traditional networks to provide high-reliability voice transmission and better coverage.

Some Key Findings

• Due to limited spectrum allocations to carriers, wireless systems- especially wireless data systems- are nearly always bandwidth constrained. This means that channels spend a significant percentage of time at full capacity, making QoS capabilities paramount.
• Wireless channels operating in the real world experience widely varying properties. Next-generation technologies raise capacity in part by operating near the Shannon limit for the channel through these varying channel conditions. This raises the overall capacity of the channel but introduces higher variability in bandwidth, which can wreak havoc in poorly implemented networks.
• The use of a shared data channel, especially when combined with the limited and variable channel conditions described above, provides opportunity for users to significantly impact one another' s user experience. The VoIP application is more susceptible than most others to the effect of a shared channel at maximum capacity.
• The delivery of VoIP services on a wireless channel can strain the capabilities of the wireless infrastructure and user equipment by tying up additional processing power than would be needed for non-voice applications.
• Next generation wireless technologies- WiMAX in particular- have done little to ensure that VoIP over wireless solutions from vendors are adequate and consistent. There are no real guidelines for packet classification or the application of a Grade of Service to VoIP packets, meaning that carriers can find themselves on their own as they try to architect their VoIP wireless network.
• The current number of worldwide Vo4G subscribers is less than 300,000 (most of which are VoWiMAX) and an estimated 20% of the cumulative total of WiMAX subscribers.
• 86.5% of worldwide VoWiMAX subscribers are in the 3.3-3.8 GHz band.
• Europe leads the number of Vo4G subscribers with a share of 35%, followed by Asia Pac with 25%.
• 68% of all VoWiMAX subscribers are getting services based upon the 802.16- 2004 standard. Iberbanda, Irish Broadband and ZAIN are the major contributors to the global VoWiMAX subscriber base with services based upon 802.16-2004, whereas Wateen and Mobilink are the major contributors for 802.16e-2005 VoIP services.
• There are significantly fewer options in LTE for vendors to differentiate than there are for WiMAX, which will likely provide carriers with a well-defined roadmap for deployment. In order to better facilitate mobile voice, LTE has been defined with a much shorter frame size to reduce delay- 1ms vs. 5ms in 802.16e-2005- and an uplink modulation that maximizes range and battery life.

Questions Answered By This Report

• What is driving VoIP in the latest generation of wireless standards?
• What are the contributing factors to VoIP quality?
• Why is good quality VoIP so difficult on wireless networks?
• How can carriers maximize VoIP quality on wireless networks?
• What methods are available for providing Quality of Service for VoIP?
• What aspects of VoIP support are indicated by the standards and what are left up to the vendor?
• What are the impacts of dynamic vs. static VoIP bandwidth assignment?
• What is VoIP call admission control and how can it improve the user voice experience?
• What are the high-level differences in VoIP support for fixed WiMAX, mobile WiMAX, and LTE?
• How are the top fixed and mobile WiMAX and LTE vendors currently supporting VoIP?
• What is the status of current VoIP over 4G deployments?
• What is the expected ARPU generated by wireless VoIP?

Who Needs This Report?

• 4G Operators
• WiMAX/LTE Chipset Vendors
• WiMAX/LTE CPE Vendors
• WiMAX/LTE Device Vendors
• Handset Vendors
• Consumer Electronics Manufacturers
• ASN Vendors
• Regulators
• Other