Spectrum fragmentation in the rush to LTE

15/02/2012
Article

Only a few years ago it was predicted that the multiple standards in the analogue days that gave way to two main camps in digital would soon become one big happy family once LTE takes hold and Wimax is quietly buried and forgotten.

However, the rush to forth generation networks and the need to provide wireless broadband access has created a maze of frequencies that have shattered the idea of homogeneity.

The other day I was sitting down reminiscing with John Stefanac, head honcho at Qualcomm for Asia-Pacific. He asked me to guess how many active LTE profiles there were out there.

Qualcomm makes chips for handsets and if anyone is to have a feel on the network rollouts across the region, it would be this company.

“Ten?” I guessed, sheepishly. Active being the imperative word.

“No, forty-three,” came the answer. I wish he had said forty-two as that would have been the ultimate answer to life, the universe and everything.

Just as soon as quad and penta-band 3G phones were becoming the norm, at least at the high end, along comes LTE and suddenly Qualcomm is expected to support 43 different frequency profiles across the globe.

Stefanac was quick to point out that there is one thing for the chipset to support a profile, and another for the handset maker to implement it. The 850-MHz/900-MHz 3G split is a case in point. For ages, nearly all Qualcomm chipsets have supported both bands, but radio limitations meant that most handset makers chose to only implement one low band in their design.

For the consumer, this has led to a situation where many handsets are sold with 900-MHz or 850-MHz variants side-by-side and makes the hard-worked-for network neutrality and no-locking policy of some regulators a moot point.

Fast forward to LTE and it will be impossible for all 43 actively deployed profiles to be supported in one tablet or phone. Quite how that will translate to the device ecosystem is anyone’s guess. Probably a return to dongles and Mi-Fi type mobile hotspots rather than built-in radios for all but the most popular profiles.

It has also been suggested that future business notebooks will see a return to thicker bezels just to house the myriad of antennae needed.

I posed the same question to Huawei’s CTO for Asia, Michael MacDonald. As a device maker (which has set its sights on becoming world number three by 2013), I asked him to name the bands where there is most traction when it comes to LTE devices.

Huawei sees LTE networks falling into two markets. The natural migration bands such as 1,800-MHz and 2.1-GHz where they will co-exist with 2G and 3G and new bands.

He also stressed that the bands that many networks were launching today are not the bands that they plan to use long term. For instance, M1 in Singapore has launched on one set of bands (1.8-GHz and 2.6-GHz) but their long-term plans are to transition to 2.1-GHz, he said.

That said, MacDonald said that 2.5-GHz, 2.6-GHz, 700-MHz and AWS (1700-MHz up, 2100-MHz down) are very strong.

Japan, China and Korea, which are the biggest markets in this region, are focused on 2.3-GHz, 2.5-GHz and 2.6-GHz. The Philippines wants synergy with North America and thus it is focusing on 700.

Well, he mentioned seven profiles, which is a bit less daunting than Stefanac’s 43.

While it is generally accepted that going against standards would be futile, there is the option when using frequency division duplexing profiles whether to go with standard paring or adopt an ad-hoc paring that is within the specifications but not expected.

Many state enterprises have suddenly found themselves in possession of such chunks of free spectrum that just fit into some profiles.

Most vendors say that while it should work, the risk lies in validation of equipment and also in destroying the value of spectrum if later, when spectrum is re-farmed, those non-standard parings remain.

Related content