LTE, or Long Term Evolution is being widely acclaimed in the media as the new messiah to bridge the digital divide. Don’t get me wrong - LTE is a good thing. But there’s a lot to do before it becomes the rural silver bullet. Find out more after the jump.
I’ve just spent a very interesting afternoon with policy officials in the Scottish government’s Innovation and Digital Economy team talking about the challenges (topographical and economical) regarding the delivery of next generation access. We of course talked about the fact that FTTH, even in its GPON variation will not be deployed everywhere - the cost and incentive simply do not add up.
With the UK government allocation of £830 millon, and with BT investing £2.5 billion to cover two thirds of the UK population, this leaves us around £25 billion short. So of course, there’s a readily accepted reality check that we have to look at alternatives to solve the rural broadband problem. We talked about satellite, which doesn’t deliver super-fast broadband (more on this topic in my previous blog post, Sky rockets in flight ? It’s time to get real about satellite internet broadband) but the discussion quickly crossed over to 4G mobile broadband, and more specifically around LTE, and the number of trials and deployments that have been announced recently.
Much as I’m a big fan of mobile broadband and where it’s heading, I think we have to be cautious about making it the silver bullet for rural broadband connectivity issues. LTE certainly has the potential to do so, but this depends on a lot of factors, notably: cost of spectrum, MNO priorities and ROI targets, and backhaul capacity.
Cost of spectrum for LTE is a key issue. If we look at the €1.8 billion reserve price that the French government has set for the lucrative 800 MHz spectrum, this comes down to €60 million per MHz (!). Even the less attractive 2.6 GHz spectrum is being set at €700 million (€10 million per MHz). It becomes quite clear that MNOs are not going to hurry deploying LTE base stations in remote and rural areas first. They’re going to deploy those networks where the adoption rate is likely to be highest: densely populated urban areas, with affluent users able to pay the required tariffs and buy the latest generation LTE handsets / dongles (which is another issue altogether, as cost effective handsets are going to take a while to appear). If you look at Verizon Wireless’ LTE network, this is a case in point: all initial markets planned are in significantly populated areas. This is not so different from the Fibre to the Home debate. FTTH will get deployed in areas where the fibre homes passed rate is highest first. I postulate that LTE will get deployed in much the same way (i.e. in areas with higher population density), unless operators take a similar approach to the one taken by VZW in the US, which is to provide spectrum usage to local, regional wireless operators in rural areas. However, this comes with a string attached. The rural LTE base stations have to be backhauled at the onus of the local operator, and herein lies the second problem: the backhaul conundrum.
LTE is a radio access techology. It provides reasonably fast access speeds (12Mbps down, 2Mbps up), but if the pipes behind the LTE base station are too narrow, then all the benefits of such a technology go out the window. LTE requires a lot of underlying capacity to deliver on its rural promise. So backhaul is key. For the ultimate experience, this will require fibre based backhaul (fibre to the base station), bringing us back full circle: fibre will roll out to those base stations where the operators ROI will be highest. So how do we solve this conundrum?
For MNOs to deliver cost effective and pervasive (i.e. urban and rural) 3G and LTE based mobile data services, they have to successfully manage and reduce their input costs if consumers are to enjoy pervasive and affordable services.
A major input cost for MNOs when delivering mobile data capacity is backhaul capacity for base station sites (see figure 1 below) especially as average consumption of mobile data increases from circa 1-2GB per month to the level of fixed broadband consumption of circa 10-20GB per month. This increase in consumption has a knock on effect on the typical requirements for a MNO cell site which has risen from circa 2MBps for a 3G site to circa 100Mbps for a LTE site (see figure 2 below).
To keep this OPEX cost to a minimum, point to multipoint (PMP) microwave solutions at bands such as 40GHz can be used. The 40GHz band offers 3GHz of spectrum for transport purposes. Available wide band PMP systems leveraging these frequencies can aggregate traffic to better utilise high capacity backhaul fibre, essentially extending its reach. In independent analysis from Informa1, it is estimated that a tier one MNO operating across five major European countries could save circa Euro three billion over a five year period. Informa estimates that PMP microwave solutions can be deployed at 5% the cost of fibre and are 40% more cost effective to deploy compared to Point to Point (PTP) microwave.
In Figure 3, we compare a Point to Multipoint (PMP) network architecture with a Point to Point (PTP) configuration. As can be seen, a PTP configuration requires dedicated links to each node served. This contrasts with the PMP architecture where a central station serves a coverage area and multiple end stations with in that area. The key cost drivers for either type of network are:
In summary, for a typical MNO network deployment, a PMP architecture can support a flexible network topology which can result in circa 65% CAPEX savings and circa 80% OPEX savings compared to an equivalent PTP backhaul network (excluding any spectrum fees from the comparison).
So what can we conclude from this? LTE has a true potential to deliver pervasive bandwidth in rural areas. However, regulators and operators cannot afford to ignore the implications on the backhaul side of the equation. The network is only as strong as its weakest link. Of course, this will be obvious to some of our readers, but I hope that this post serves as a little reminder that the key to successful LTE deployments will revolve around rapid spectrum allocation of not only 800MHz and 2.6GHz frequencies, but also a differentiated approach in approaching the backhaul conundrum, especially when targeting rural areas, where the priority of deploying fibre to the base station is not very high.
Usage of widely available, unallocated spectrum in the 40GHz bands for example, can bring to the market innovative, high capacity point to multipoint (PMP) solutions for the access and backhaul segments. These solutions can significantly reduce costs for MNOs to deliver 4G LTE services using bands such as 2.6GHz and 800MHz, which in turn can significantly reduce market pricing and (by allowing for an enriched alternative to fixed services) enable consumers in rural and remote areas to receive high quality broadband services.