In his wireless blog, Alastair Brydon laments the continuing lack of long term solutions for voice over LTE, as Verizon and Huawei announce products that depend on legacy networks.
I notice that 2011 has begun with announcements of “breakthrough” developments that enable LTE users to make voice calls. Given that the early launches of LTE have been focused initially on data services, these announcements immediately grabbed my attention. Unfortunately, it turns out that neither of them allows voice to be carried on LTE. In both cases the solutions use “old” technology to carry the voice call. While this allows operators to exploit their legacy investments, it concerns me that it may be many years before legacy spectrum can be refarmed to use much more efficient technology, such as LTE.
In the USA, at CES 2011, Verizon Wireless announced the launch of four LTE handsets, from HTC, LG, Motorola and Samsung. The handsets will allow users to undertake voice calls while simultaneously connected to the Internet, for example to discuss the content of a web site. However, this is not achieved by an elegant integration of voice and data onto the same LTE connection. Instead, the phones contain dual receivers, which enables them to establish a voice call on (2G) cdma2000 1xRTT in parallel with an LTE data connection. It is more than likely that such an approach will have significant implications for battery life and it is notable that the first shipments of these devices will actually have this feature switched off.
In China, Huawei has announced that it has successfully achieved circuit-switched (CS) fallback in an LTE network, using a third party mobile on its own LTE/EPC infrastructure. This feature, which is standardised by 3GPP, allows a mobile to leave LTE to establish a call on an underlying GSM or UMTS network, if a user requires voice telephony. Huawei described this a “significant” development, which allows mobile operators to benefit from their legacy investments. It is true that this adds a crucial feature to LTE networks that are inherently designed for data only. However, it is a very odd state of affairs that sees the introduction of LTE as a breakthrough in wireless technology, which is unable to carry the most valuable mobile service of the last 25 years.
From the mobile operator viewpoint, there may be commercial benefits in tying voice telephony to their legacy networks, where they can keep a tight control on its delivery and limit the opportunities for competitors to displace them. In contrast, the much more open IP-architecture of LTE may present easier opportunities for others to compete with the traditional mobile operators for the delivery of mobile services. However, from the point of view of the efficient use of resources, it would be better to move voice traffic to LTE as soon as possible. This is not because of any deficiency in the ability of legacy technologies to deliver voice telephony. Indeed GSM networks do a fine job of delivering voice services worldwide, often with much superior coverage to the corresponding 3G networks. The issue is that there is a major missed opportunity in releasing 2G spectrum for use by LTE.
As we have discussed in previous posts, the rapid growth in broadband data services puts spectrum at a premium in the wireless industry. Putting aside the issue of investment in legacy technology, it is clearly beneficial to deploy the most advanced technologies in the available spectrum, in order to maximise the benefits of that technology. Substantial improvements have been made in mobile technologies over the years. For example, we estimate that the spectral efficiency (i.e. throughput per MHz of spectrum) of LTE is about 25 times that of GSM. Therefore, it is important to encourage the rapid migration of spectrum from 2G to next-generation technology, as illustrated in the figure below, taken from our response to the UK government consultation on spectrum modernisation.
The figure shows the effective downlink bandwidth available, to a typical UK mobile network operator, relative to deploying HSPA in the same bandwidth, and how this could change over time as LTE is deployed in each frequency band. It shows that 2G technology contributes relatively little to total capacity today (despite having a substantial amount of spectrum allocated to it), and refarming of this spectrum would provide a significant uplift.
One factor that will inevitably slow the refarming of 2G spectrum is the need to support 2G mobile users for as long as they continue using 2G networks. Rather than depending on 2G for the delivery of voice services, mobile operators need to encourage their customers to migrate away from 2G services as soon as possible, so that they can subsequently turn off their 2G networks and use the spectrum for LTE.
For some time there has been work in progress to provide medium- and long-term solutions for the delivery of voice over LTE. The VoLGA Forum is promoting an approach based on the 3GPP Generic Access Network (GAN) standard, which transforms the cellular voice service into mobile voice over IP delivered via LTE. Meanwhile, the VoLTE group of operators, supported by GSMA, is promoting a solution based on IMS. However, there are no signs of either of these technologies being implemented soon, and the deeper any short term solutions become embedded in LTE networks and handsets, the longer it will take to evolve them out. For the remainder of 2011 I hope to see much more progress on these longer term solutions.