AT&T will deploy 40,000 small cells - if they can find the sites!

by Terry Norman

Over the past number of days I’ve been attending the fantastic Small Cells World Summit 2013, in London. It was a great insight into operator deployment strategies and the challenges they have to confront in this space. An oft chanted mantra was ‘getting the right site is key’. However, a series of presentations showed that this is not trivial. Before we go on to consider why, we first need to be clear on the characteristics of a site which makes it suitable for a small cell.


AT&T plans to deploy 40,000 small-cells (in addition to its home femto cells) as part of the massive densification program over the next 5 years to support a projected tenfold growth in data traffic. When dealing with deployments numbers on this scale, acquiring the site needs to be operationally cost-effective. In addition placing the site at the centre of traffic is critical. AT&T discussed the importance of what it termed the surgical placement of the small cell to maximise its usefulness. Similarly, Sprint, in its presentation, showed that the cost-benefits of the small cell was most sensitive to the location of the cell - and thereby its ability to support clusters of users at the centre of the traffic.

So, to return to our earlier question: what makes a suitable site?

  • The small cell site needs to be placed where users gather:  Rob Joyce, taking time from his sabbatical from Telefonica to complete his PhD thesis at Leeds university, points to how difficult it can be to locate the centre of the traffic. For example, inaccuracies with GPS mean that the centre of a traffic cluster is effectively blurred. Indeed this is the motivation behind Rob's research which is developing algorithms which use both GPS and network metrics to accurately locate the centre of the traffic.
  • Acquiring the site must be straightforward: However, assuming the operator is able to define the traffic centre- or at least take an informed ‘guess’, the next problem is' the acquisition of the site. Local government regulations, contractual negotiations with the site owner, permits all present their own operational challenges which increase operational cost and distract the operator from its day to day business of delivering mobile services to its customers.

There are other important features which define a site’s suitability, like power at the site and backhauling should be straightforward, but these are almost secondary to the challenges of defining the centre of traffic and acquiring the site. And the coincidence of these two requirements presents the MNO with a not inconsiderable challenge, especially when deploying small cells in the large numbers required making them an effective solution.

These tough challenges are driving industry interest in SON for small cells

As a consequence of these challenges, we find operators are taking the path of least resistance and deploying the small cells to locations where they already have easy access – where they own the rights of deployment, or can acquire them wholesale through negotiation with vendor chains or local governments as examples of indoor and outdoor deployments, respectively. AT&T is taking this low-hanging fruit approach, at least in the early stages of its densification plan. It will deploy small cells to locations where it has easy access, compromising upon the ideal of acquiring a site at the centre of the traffic cluster. Given the likely inaccuracy of any guess on where the traffic centre is, this is a sensible and pragmatic approach, but of limited opportunity. There are only so many sites like this that are available, and at some point AT&T may have to come back to plug the gaps – let’s hope the self organising nature of small cells by that time will be good enough to cope.

We expect other MNOs that have the financial muscle and presence to follow the pragmatic approach, at least over the next two years. They will be hoping that SON (Self Organising Network) capabilities will mature sufficiently for them to sew small cells into the gaps which remain. We hope so too, otherwise network densification using small cells could become a very painful, costly process for operators.

As a footnote - the arrival of new layer of small cell intermediaries, whose role is to acquire, deploy and manage the small cell sites, is testimony to the difficulty of dealing with the confluence of these two requirements – defining traffic and acquiring the site. This neutral host model may have growth potential. Already there are companies seeding this fertile ground, for example Cloudberry – offering Small Cells as a Service - operating out of Norway and UK. Wireless Explorers will be keeping a keen eye on developments in this space.