Your browser is not supported. Please update it.

6 February 2008

Giants and minnows adopt strategies to feast on LTE transition

It is always the way ‘ threaten a hugely profitable business like cellular and suddenly it finds it has plenty of political will and cash to up the stakes and take on any newcomers. That’s pretty much the impression we have had since LTE was on the distant horizon and WiMAX mobile began to become more concrete. Suddenly LTE is looming in the WiMAX mirror, coming up on the outside.

Of course LTE has become closer and closer to WiMAX and the technologies are more similar than anyone expected them to be, which makes it easier for vendors that support both and especially those that took an early lead in WiMAX. However Ericsson isn’t one of those, it is merely the company that has most to lose by the world switching to WiMAX.

So in the past couple of weeks we have seen LTE propelled by Ericsson into the TDD profiles traditionally associated with WiMAX, while the WiMAX Forum’s recent summit made progress in the reverse direction, towards FDD implementations of 802.16e. And UK-based picoChip, working with German start-up MimoOn, has produced the first full reference design for LTE base stations.

The picoChip designs, which range from femtocells to macro base stations, take advantage of the commonalities between 802.16e and LTE, much touted by equipment vendors working both technologies (such as Motorola, which has found an overlap of over 80% in its development of infrastructure for the two platforms). The LTE products use the same physical platform as picoChip’s WiMAX family, enabling the start-up to steal a march on larger semiconductor makers ‘ with a reference platform that, in turn, VP of marketing Rupert Baines sees as reducing time to market for smaller equipment vendors, as they seek to use new technologies to break into the tight-knit club of GSM and UMTS base station makers.

The PC86xx LTE reference design was codeveloped with MimoOn, a German start-up focusing on IP for MIMO, in a non-exclusive agreement. Like its WiMAX stablemates, it uses picoChip’s own PHY and the Wintegra MAC, and. The PHY runs on the picoChip PC203 devices and includes OFDMA downlink and SC-FDM uplink, with support for up to 4×4 MIMO and for AAS, and supports TDD and FDD modes. MimoOn provides an end-to-end full software implementation of LTE on multicore DSP platforms like the picoArray.

We’re not sure that competition from smaller companies will in fact materialize in LTE, because of the immense market control exerted by the traditional cellular equipment makers, but it seems that even greater efficiencies can be achieved if a platform can be adaptable must by changing software, between more than one technology. The likes of Intel and Ericsson may be pitting WiMAX and LTE directly against one another, but in reality most suppliers will want both, and want them as cheaply as possible.

Among the tier one base station makers, only Ericsson is strictly LTE-only (and even the mighty Swede OEMs Airspan kit for customers in DSL-style applications); Motorola and Nortel have majored on WiMAX, having failed in UMTS, and want to reuse that work in LTE, and Samsung may do the same; Alcatel-Lucent, Nokia Siemens and the Chinese majors are all working on both platforms.

Another factor that will make LTE disruptive rather than conservative will be the emergence of new base station formats, such as femtocells and broader use of picocells, which will support new operator models, and may also enable non-traditional equipment makers to elbow into the territory of Ericsson et al. ‘Ericsson’s world will never be the same again,’ bemoaned one senior executive last week, referring to the Swedish giant’s belief that ‘ however painful for its traditional profit model ‘ femtocells will become a commercial force from 2009, and mainstream from 2010, similar timescales to those expected for LTE and Mobile WiMAX.

With these factors in mind, reference designs need to be flexible, and the PC86xx from picoChip scales from single-chip femtocell designs to multi-sector macro stations, from 1.25MHz to 20MHz and with both TDD and FDD modes supported. Being software defined, it can be upgraded in line with refinements to the LTE standard, which is 80% finalized, and the common platform also supports Wave 2/MIMO Mobile WiMAX now, and will support upgrades to the forthcoming FDD profile and to the next iteration of the standard, 802.16m.

Meanwhile, though the WiMAX Forum has been officially coy about the FDD profile for 802.16e, it seems that work on this long-planned enhancement has accelerated in the face of Ericsson’s aggressive moves to push LTE into the TDD spectrum, which will increasingly be the focus of regulators’ new allocations, as the developed economies intensify their focus on mobile broadband data applications, to which TDD operation is well suited.

Internal documents from the WiMAX Forum’s recent meeting in Hawaii have circulated widely, indicating that the Forum’s technical working group has now received full recommendations for an FDD profile for three frequency bands. One of these, predictably, is 2.5GHz, one of the two core bands for 802.16e (along with 3.5GHz), but the others show WiMAX moving directly into the cellular standards’ territory, and also taking a highly US-centric focus, with recommendations for two bands that are mainly available in the US ‘ AWS, auctioned last year, and 700MHz, the subject of the current FCC auction. Individual vendors like SOMA and Alcatel-Lucent have been adapting WiMAX for both these bands already, but now they appear to be moving to the WiMAX mainstream, stepping up the challenge to HSPA and LTE.

Now Ericsson has also shown the first demonstration of LTE running in both TDD and FDD mode on the same base station platform, claiming major economies for operators that can use the same system in both unpaired and paired spectrum. Although many carriers, especially in western Europe, had TDD spectrum allocated along with their FDD 3G licenses, few built it out because of limited amounts, and because it required a separate network (with the TD-CDMA technology from IPWireless, now owned by NextWave, the only real product option. However, IPWireless’ early R&D on a TDD path to LTE may put its new parent in an interesting position in terms of future products or licensing opportunities). Now Ericsson claims to have overcome that problem and its demonstration showed LTE in TDD mode working at 90Mbps in the downlink with 2×2 MIMO. Ericsson has previously demonstrated LTE in FDD mode with speeds of up to 160Mbps, as have Nokia Siemens and other key vendors in this market.

Ulf Ewaldsson, VP of product area radio at Ericsson, said: “We have now demonstrated that LTE can be used in both paired (FDD) and unpaired spectrum (TDD) on the same base station platform. This means that operators with only unpaired spectrum also can plan for LTE and benefit from the optimal global LTE solutions.”