Your browser is not supported. Please update it.

28 February 2013

A few cellcos make LTE-A real, but let’s not discuss 5G just yet

As predicted, there was intense focus at Mobile World Congress on LTE-Advanced, even before the vast majority of carriers have deployed basic LTE to any great extent. Also as expected, there was a degree of over-hype about LTE-A, and particularly a tendency to portray it as a big-bang upgrade, rather than a menu of options which can be used to enhance existing LTE performance. On that menu, much the most popular elements to be launched – by chipmakers, equipment vendors and a few brave cellcos ‘ were carrier aggregation, which will start to creep into real services later this year, and self-organizing functions which will be essential to small cells, such as automatic neighbour relations based on the X2 standard.

While suppliers showed off roadmaps and a few commercial products, all eyes were on the early deployers of LTE-Advanced features. As in most of the LTE story so far, the laurels were seized by the Asian giants, particularly SK Telecom and NTT Docomo, both of which announced extensive roll-out plans and vendor relationships.

These will be joined this year by Australia’s Telstra, by AT&T and other US players, and a few others in early adoption of at least a few elements of the diffuse standard. But for now, Docomo and SK Telecom were setting the pace. The former announced the vendors, Nokia Siemens and Panasonic, which will upgrade its network with functions like carrier aggregation.

This is a good win for NSN, which has not featured as heavily as Ericsson and Alcatel-Lucent in the most advanced LTE roll-outs to date. Breaking into the Japanese carriers is tough, since Docomo in particular tends to rely on trusted local suppliers with which it has long-standing development alliances.

Panasonic, of course, falls into that category ‘ it has worked with the operator since 2007 on LTE network infrastructure, but NSN was also brought into that project at an early stage and its efforts have borne fruit. The European vendor will supply its Liquid Radio multiple standard RAN. As in Docomo’s first stage LTE network, there will be heavy use of remote radio heads, with baseband processing virtualized in the cloud, as well as increasing roll-out of small cells to increase indoor and outdoor capacity. NSN says the base stations will deliver capacity of 300Mbps.

In future, the two vendors will support Docomo’s own particular definition of Cloud-RAN, a concept which is being pioneered in China, Japan and South Korea, and which takes the idea of remote radio heads and shared basebands to a new level. Docomo says it favors C-RAN because the cell site equipment, consisting of radio and antenna, is compact and low power, and so can be deployed quickly in high traffic areas like train stations. It calls its architecture Advanced C-RAN and this will rely on some HetNet principles, including a separate layer of ‘add-on’ small cells adding localized capacity while cooperating with macrocells.

In the C-RAN, there will be high capacity master base stations supporting multiple macrocells plus the local small cells. The master BTSs will handle up to 48 macro and small cells at launch and more later. Carrier aggregation will be supported for cells served by the same base station. The carrier says it will boost peak downlink speed to 3Gbps over time, hitting ‘true 4G’ and 3GPP LTE-B standards.

One of Docomo’s interests in LTE-A is to support carrier aggregation in a standardized way. Although CA is an old concept, 3GPP Releases 10 and 11 (LTE-A) provide standard ways to combine frequencies in the same or different bands, to create channels up to 100MHz wide. The Japanese cellco recently acquired 700MHz spectrum and so will have three LTE bands as well as its 3G spectrum, so its interest in CA is rising.

With its multiband situation in mind, it has also announced that it has developed its own base station antenna to support all five of its frequencies, including the newly allocated 700MHz (although that will not be freed up for commercial LTE use until 2015). The new antenna is no larger than Docomo’s existing four-band antennas for 800MHz, 1.5GHz, 1.7GHz and 2GHz but the firm said adding the extra frequency was challenging because mutual interference increases as bands are added.

SK Telecom is also working with NSN, evaluating the latest element in NSN’s ‘Liquid’ portfolio for modern, flexible networks ‘ Liquid Applications, which aims to support personalization by turning base stations in to local hubs for service delivery, close to the end user. Under a memorandum of understanding signed at MWC, SK Telecom will put NSN’s Radio Applications Cloud Server (RACS), which enables Liquid Applications, in the first half of 2013.

This approach is becoming increasingly important in LTE as operators seek to deliver differentiated levels of quality of service to different subscribers, as well as tailored applications. Localized content delivery can also reduce the load on the overall network. Intel was vociferous about the Smart Cell technology it has developed with small cell specialist Ubiquisys, which involves incorporating an x86 processor into a base station or as a companion platform, to handle far greater quantities of local processing than a conventional BTS allows. And Samsung has a similar approach, going up to larger cells too, with its Smart LTE platform.

In NSN’s case, Liquid Applications exposes applications and services to real time network data to improve service quality and targeting. Marc Rouanne, head of mobile broadband at the vendor, said it ‘redefines the role of the base station and will transform the mobile broadband experience. The beauty of Liquid Applications is in the simplicity of using information that has always been there in the network, to fundamentally change the definition of personalized service and shatter forever the perception of the network as ‘just a bitpipe’.’

Ambitious words, but given a credibility jolt by the endorsement of SK Telecom. NSN will welcome a strategic role in the carrier’s LTE evolution ‘ it was selected, along with Ericsson and Samsung, as a 4G network provider in 2011, but in RAN terms has been somewhat overshadowed by the other two. As in many markets, it is pushing the flexibility of its Liquid network to promise operators a system that can adapt to changing requirements and traffic patterns on a constant basis ‘ and hoping that will help it make up for a relatively slow start in conventional 4G RAN sales.

SK Telecom will be the first operator to test Liquid Applications in an LTE environment, as part of its Service Aware RAN project. Service-Aware RAN aims to inject significant intelligence at the edge of the network to deliver customized services and increase the value to users and content partners. ‘Telecom providers are realizing that optimized transmission of data alone can no longer guarantee success in the market,’ said Dr Choi Jin-sung, head of SK Telecom’s ICT R&D division, while Jarmo H’kkinen, head of mobile broadband value creation at NSN, added: ‘Liquid Applications will stimulate mutual cooperation between the operator and various ecosystem players, resulting in development and delivery of innovative applications and services. This will help SK Telecom to generate revenue from new business models, new services and new applications.’

NSN had to share the SK Telecom limelight with its fellow RAN suppliers however. Both Samsung and Ericsson were showing off the advances they were delivering to their star Asian customer, and the Swedish giant signed an MOU of its own, for cooperation on LTE-Advanced, and particularly a technology called SuperCell 1.0. The collaboration will focus in particular on small cells and the two companies demonstrated LTE-A SuperCell 1.0 ‘ designed to almost eliminate inter-cell boundaries and make groups of small and macro base stations act as a single cell by allocating identical cell Ids, using another item from the LTE-A menu, CoMP (coordinated multipoint). This removes the need for handover, supporting faster data rates even at the edge and a more seamless user experience, and unlike carrier aggregation in its current standard form, it boost uplink speeds, not just downlink. SK Telecom plans to apply SuperCell 1.0 to its LTE-A network in the second half of this year.

On the device side, SK Telecom said it is in talks with multiple device makers for LTE-A handsets but has not made selections yet. Most urgently, it wants smartphones that will support carrier aggregation across its 850MHz and 1.8GHz bands, which could boost its peak LTE speeds to 150Mbps on the downlink (though CA does not currently help with the uplink).

Many chip suppliers have launched the Category 4 modems which would be necessary for that data rate, and Huawei has claimed the first commercial handset with the capability, the Ascend P2. ‘We plan to launch our first CA devices in 2013,’ said Chanwoo Park, VP of Samsung’s mobile product strategy. ‘3G has four major frequency bands, LTE has more than 20, and they are being used in more than 22 different combinations. That makes it very hard to support all the customers around the world with one device.’

Another operator which is always at the cutting edge of new networks is Telstra, and the Australian incumbent was also announcing expanded LTE-A cooperation with Ericsson. The two firms agreed to expand the carrier’s 4G LTE network across Australia and to start planning for LTE-A, initially in the form of carrier aggregation across 900MHz and 1.8GHz bands. Other plans include a trial of LTE HetNet and another of LTE Broadcast.

Ericsson CEO Hans Vestberg said: ‘Under this agreement, Ericsson continues as Telstra’s prime vendor of 3G and 4G RAN technology. The trials for Telstra will provide insights applicable not only on the Australian market but also globally. For example, the heterogeneous networks trial will be of interest to operators around the world who want to ensure capacity and connectivity in densely populated city areas. And the LTE Broadcast trial will show how Ericsson technology can deliver media content efficiently over LTE.’

These operators may be in the small minority of cutting edge deployers of LTE-A, but their roadmaps are very real and firmly grounded in business goals. However, some suppliers are looking further into the blue yonder, with talk of LTE-B, ‘5G’ and other terms which are hardly defined in the industry. Indeed, the 3GPP urged caution about such labels. It said there was no timescale for the next major architectural shift ‘ which would be the appropriate time to introduce the 5G label ‘ as there was still much more to be got out of the current OFDM-based platform that underpins LTE. And while it is starting the process of defining its Release 12, which some are calling LTE-B, it does not plan to use that nomenclature ‘ and even if it did, the standards will not be frozen until late this year.

That does not stop companies claiming LTE-B compliance based on pre-ratification standards work, or just on assumptions about what the specs will contain. Huawei unveiled FusionNet, which it described as a ‘new generation of network architectures for LTE-B’. The platform does sport some very advanced technologies which will be highly relevant once carriers’ start to deploy small cells and HetNets on a large scale. It promises a fivefold improvement in cell edge user throughput by harnessing some of the innovations which are known to be part of the Release 12 work. But that does not mean that using such a term is not misleading, implying a standards-based approach which is not yet possible.

FusionNet is still a concept network, to be fair, and an interesting one. It builds a multi-access, multiband and multi-layer HetNet using several core elements – multistream aggregation, interference coordination, traffic adaption, and spectrum efficiency optimization. It coordinates existing LTE and LTE-A components like CoMP and carrier aggregation and applies them more integrally to a dense network, delivering more universal service and reducing cost.

The Chinese vendor also unveiled its Ultra Node prototype design for future base stations, promising an eventual peak data speed of 50Gbps in one node. This is achieved using a redesigned antenna structure, multistream aggregation of ultra-widebands, and greater channel capacity via an advanced multi-antenna encoding algorithm. Some of this may find its way into proposals for future 3GPP standards since Huawei has been a prolific contributor (362 proposals for LTE standards, more than any other company, it says).

Huawei may be looking into the blue sky with its R&D, as the balance of innovation and IPR collection shifts to Asia, but Neelie Kroes, the European Commissioner for the digital agenda, wants to slow that shift and put Europe back in the mobile technology cat-seat, where it sat in the 3G era. She plans to invest heavily in future wireless R&D, though she hardly helps the 3GPP’s anti-hype cause by labelling this activity ‘5G’. Kroes is committing €50m ($65m) to ‘5G’ research and has set a 2020 goal for the ‘delivery’ of the technology shift.

The new funding includes €16m for the METIS project which is looking at faster mobile broadband technologies which boost power and spectrum efficiency. The vendors involved in that project include operators Deutsche Telekom, France Telecom, Telefonica, Telecom Italia and Portugal Telecom, plus vendors Alcatel-Lucent, Ericsson and NSN), and academic institutions.

‘Europe used to lead the world on wireless. European 5G is an unmissable opportunity to recapture the global technological lead,’ Kroes said. ‘Rolling out today’s networks is important. But what comes after? For the next global standard, and the next generation of technology, will Europe lead the world, or merely follow?’ Judging by MWC, where those in the vanguard of 4G evolution were heavily concentrated in Asia, she has a tough mountain to climb to restore Europe’s lead.