The cable TV companies have been in the news a lot recently, with some significant decisions announced that change or reconfirm industry direction. The past week has seen major announcements from Comcast, Time Warner and Cox communications as they release spectrum and wireless network assets. Meanwhile Sprint made some major changes at Clearwire (which they own 50% share of), while Verizon is well placed to take advantage of closer ties with the cable companies. Sprint plans to use Clearwire as part of their LTE solution - we ask if this is the most cost effective approach and consider the implications.
US Cable companies withdraw from their own cellular services
A few years back, several US Cable companies had made significant (multi-billion dollar) investments in wireless. Comcast was one of the investors in Clearwire, the WiMAX network heavily backed by Sprint, Intel and others. Comcast invested $1.2 Billion into the Clearwire WiMAX venture and still owns about 9%, but after writing down $600 million has indicated that it is unlikely to increase its investment.
After a foray into building out their own mobile network in the US, Cox Communications decided this month to exit this part of their business. Citing a lack of scale – Cox have some 6 million wireline/cable subscribers – they had found it difficult to compete with larger wireless networks that could invest more in higher speed networks (including LTE) and offer a slicker range of handsets.
The spectrum that Cox together with other licences held by a consortium including Comcast and Time Warner has been sold to Verizon Wireless for $3.6Billion.
Change of direction for Clearwire too
For Clearwire, the US WiMAX operator, this doesn't look like good news. Still not profitable, they could also be said to be suffering from the same generic problems as Cox – a lack of scale. It takes very deep pockets to build out a brand new network and continue investing before becoming profitable. With Sprint making a major technology shift and announcing major investments in LTE, promising commercial service next year and LTE Advanced in 2013, Clearwire will also deploy LTE next year.
Sprint have agreed to invest a further $1.6 Billion with Clearwire, who will deploy TD-LTE in high traffic areas while continuing their WiMAX service until at least 2015. This will complement Sprint's own LTE rollout, providing extra capacity in these potential problem areas. Although it sounds a lot of money, this investment will be spent over 4 years and is considerably less than what Verizon and AT&T are committing to. Clearwire aren't yet saying which cities they'll build out first.
This is in addition to Sprint's previously announced wholesale deal with LightSquared, a wholesale LTE network which plans to reuse it's satellite spectrum but under threat because of interference concerns with GPS systems.
Although Clearwire have plenty of spectrum (some 120MHz in many parts of the US), it's at quite a high frequency (2.6GHz) which gives much shorter range than the 700MHz used by Verizon and AT&T. This means more basestations are required to cover large areas, but in dense urban areas with high traffic this large amount of spectrum helps add capacity. Clearwire have more than enough spectrum to build and operate both WiMAX and LTE should they wish, but will need a lot more investment to scale up to match the existing US mobile operators.
Sprint will use Clearwire LTE for capacity offload in hotspot areas
With the level of investment committed so far, Clearwire will target their coverage on those areas of high traffic demand where Sprint's LTE network could be overloaded.
Sprint have clearly made a choice between investing in Clearwire or scaling up their own LTE deployment and using more small cells. $1.6 billion would pay for quite a lot of public access LTE femtocells. Clearwire have the ability and workforce to do build out network quickly when the finance is in place, as they've proven for their WiMAX deployment. But this will be at a different frequency (2.6GHz) and use a different mode (TDD rather than FDD) than other mainstream LTE deployments in other parts of the world.
Will lack of devices supporting TDD mode be a problem?
A Heavy Reading analyst report published March 2011 forecast a split of 1/3rd TDD versus 2/3rds FDD worldwide LTE deployments by 2015. Ovum forecast a more cautious 25% market share (February 2011). Although Clearwire/Sprint were the first to raise the request for 3GPP to standardise the 2.6Ghz/TDD frequency/mode combination in March 2010, many others around the world have also shown strong interest. It was published in the specifications in March 2011 and several networks are actively trialling and deploying it today.
Perhaps the biggest concern would be whether there will be compatible handsets and devices to operate in this mode. Undoubtedly the device vendors will concentrate on the most common frequencies and modes where possible to keep costs down and reduce risk. But if the most demanding users (by traffic volume) are using USB dongles, these typically can be deployed more quickly and cheaply than replacing more complex smartphones. It is typically these which can be offloaded to greatest effect because they generate much more traffic.
Small cells would be the obvious alternative
Investing the same money in a small cell approach would allow greater compatibility with the widest set of LTE devices. All customers would be able to benefit from the full network capacity. The cost of deploying larger numbers of small cells has been shown to be more attractive than increasing the number of macrocells.
It may be that Sprint or Clearwire has in mind to deploy LTE small cells in public areas too – that's not been publicly mentioned yet. They have pioneered CDMA femtocells for both residential and enterprise customers, being the first to launch a commercial nationwide residential femtocell service back in 2008.
Wholesale peak capacity offload?
Many industries suffer from times and areas of overload, and find ways to outsource or offload capacity to other businesses. Roaming to Clearwire LTE in peak areas is certainly one option. Building out Wi-Fi networks is another. But for my money, owning and operating your own network which has been appropriately scaled up to meet demand in these areas by careful use of femtocells and small cells seems to me to be the most attractive option.