For a mobile operator, sites are better than spectrum

(Image credit: Image Credit: Varnish Software)

For the mobile network operators (MNOs), delivering ever more capacity in their networks is a constant struggle. The usage of mobile data continues to grow at a heady rate – something like 50 per cent a year. This means that after six years the demand for data is 10x greater and after eight, it is 25x greater. Continuing to grow the network to meet such a massive increase is difficult, and not always achieved as those trying to download in busy train stations at rush hour will know well.

Yet there are some operators that are able to deliver much more data than others. Across the EU the average data usage is around 2.8 GB/month per subscriber. In Finland the average subscriber uses nearly 24 GB/month. If the average EU MNO could replicate what the MNOs in Finland are doing then they would gain 10x capacity, or around six years of mobile data growth.

The combination of Tutela’s crowd-sourced data on mobile networks, and investigative analysis from Rewheel has revealed fascinating insight into how Finland achieves these gains and why some other countries struggle. The simple answer is that Finnish MNOs have more cell sites. A denser network of sites provides a massive increase in capacity, an improvement in quality and the ability to work with less spectrum.

To expand on this further: Finland has far higher monthly usage than any other country in Europe – nearly ten times the average – yet its network delivers speeds no worse than the average and good values of latency. The high loading on the network is not at the expense of a congested and under-performing system. Conversely, German networks carry around half the European average traffic and yet underperform in data rates. Not only are their networks relatively lightly loaded, they also appear to be congested.

So what can operators do to ease network congestion? The conventional wisdom is that capacity is driven by spectrum availability. Regulators are frequently lobbied to issue new bands – such as the 5G spectrum at 3.5GHz – in order to allow for capacity growth. MNOs pay millions, or in some cases billions of dollars for licenses. But Rewheel’s analysis casts much doubt on the value of spectrum and the importance of its early release. In fact, the Finnish operators have less spectrum than average deployed on their cell sites. If anything, those with more spectrum appear to have less capacity.

In the cases of the two countries mentioned above, Finland has 3.7 sites per 1,000 people whereas Germany has only 0.7. These two examples reflect the wider trend of clear correlation between site density and data usage. Furthermore, the countries with more sites have higher signal levels, another clue to what is happening.

More base stations

Having more base stations is so important both because each additional site brings additional capacity and because signal levels are higher, resulting in better network performance. The later point is worth considering further. According to analysis of cell capacity consumed by heavy users (say downloading high definition video) compared to distance from the centre relative to cell range, capacity consumed increases near-exponentially further away from the centre of the cell range.

If a user is, say, at a quarter of the maximum range, then they consume only around 3 per cent of the capacity. At half range this rises to about 13 per cent, at three-quarters to 37 per cent and then at the edge, it jumps to all of the capacity. This is because at the cell edge the signal is weak and so the amount of information that can be encoded has to be reduced significantly. In order to send the same amount of information the users have to transmit for longer, taking up increasing amounts of resource. By putting the sites closer together, the signal levels at the edge of the cells are much higher – effectively the sites operate at half or three-quarters maximum range.

This can be seen in that Finland has five times as many base stations per person as Germany yet carries 17 times as much traffic per person and all at a higher quality. A simple assumption would be that five times as many sites leads to five times as much capacity, which proves not to be the case. The difference is due to the higher signal levels which mean subscribers send more data per timeslot and so use less of the network’s resources.

With such substantial gains in capacity from a dense network, having large amounts of spectrum is not necessary. This has particular significance when it comes to 5G planning: an operator with such a dense network would not need to acquire large chunks of the 3.5 GHz spectrum, saving money on spectrum fees. Another benefit is that indoor coverage will be improved as will the overall subscriber experience. A dense network is clearly a much better network.

That begs the question as to why operators are not making their networks denser. In fact, many are reducing plans for additional sites, hoping to use the 5G frequency bands on their existing sites as a way to resolve their current capacity issues. The answer, of course, is cost. New sites, especially in cities, are very expensive. Finding them is hard, negotiating with the landlord painful, obtaining planning permission onerous, getting power and backhaul to the rooftop challenging, rental costs are high and other national regulations may add to the pain. Governments and regulators can help with “barrier busting” teams and regulatory changes, but few have been effective at this to date. Hence, for most operators, solving immediate capacity problems is more cheaply and rapidly addressed through more spectrum, even if it is an interim solution rather than a long-term one.

Even with these concerns though, operators that do bite the bullet and invest in significant densification of their sites will gain substantial long-term benefits, with much greater capacity than their competitors, a better quality network and lower on-going costs of enhancement and spectrum access. Such an operator might be able to persuade subscribers to churn from their competitor with offerings of larger data bundles and better connectivity. In the longer term they might even force consolidation in the industry. But few operators are able to look long-term or to raise the capital needed. However, in an increasingly competitive MNO landscape, succeeding with a denser network may well be a vital opportunity for those that can.

William Webb, Director, Webb Search
Image Credit: Varnish Software