Market research, financial analysis and consulting on wireless technologies and services
Published in Mobile Europe

The growth and usage models for wireless data have run counter most of our initial expectations. The major driver to growth did not come from established telecom players— manufacturers or operators—nor from the availability of technologies like 3G, but from a surprisingly successful and disruptive market entry of two companies, Apple and Google, that until then were only marginally involved in the wireless industry.


Similarly, the expectation that wireless data growth would be initially driven by business users and business applications, mostly confined to locations where a readily available connection was not available, and following a similar time-of-day distribution as voice traffic—i.e., peaking during business hours--did not match the actual evolution of mobile data.

Instead, growth in data traffic—although to a lesser extent the growth in data services revenues—is driven primarily by entertainment applications, with video contributing to an increasingly larger portion of it. Video accounts for almost half of the traffic at Clearwire, and about a third of traffic according to Sandvine’s data collected from mobile operators. Cisco’s VNI forecast predicts that video will grow to 66% of overall traffic by 2015. While many of the subscribers are business users, the bandwidth-heaving applications on their smartphones and laptops are in most cases not work-related and their use peaks in the evening, around 9:00 pm--a time that is surprisingly consistent across cultures, countries, and operators.


Perhaps even more surprisingly, subscribers do not see mobile data as the second-best alternative to be used where wireline broadband and Wi-Fi is not available. True, the cellular connection may be slower and more expensive than the DSL, fiber or cable modem line, but it is also more convenient and in many cases faster—no need to walk to the home office to de-hibernate the laptop to check email, Facebook or Twitter.


The emerging usage patterns—prevalence of entertainment applications, increasingly dominated by real-time content streaming mostly shifted towards the end of the day— has an important consequence for mobile operators trying to accommodate this traffic: mobile data is predominantly consumed indoors, at home, in the office, or in public venues. In fact, most most subscribers are not even mobile. They are their desk, sitting on a sofa or on a bench. According to a study from Cisco, subscribers use mobile data services 40% of the time from home, 25% from work, and 35% from public locations—with at least 80% of the traffic coming from indoor locations. Jaime Lluch Ladron of Telefonica expects that 95% of data traffic will come from indoor locations in a few years’ time.

The trend towards high indoor usage at peak hour is unlikely to change, even if the mix of applications may change through time, mostly because most data applications require a stationary location, preferably protected by glare or rain. Furthermore, in the evening, the peak time for mobile internet usage, people tend to be at indoor locations.

Applications tied to social networking, email, and location-based that are more commonly accessed from outdoor locations, typically generate more modest traffic loads and they may be used frequently, but for short duration of time.
We expect to see a growth in traffic from applications that run in the background— i.e., off-line content downloads, backups and software upgrades—and this may reduce the proportion of indoor traffic. These applications are unlikely to have a substantial effect because they often run at off-peak times and, even then, the devices are most likely in indoor locations, simply because this is where we spend most of our time.


The prevalence of indoor device location when using data services has a major impact on how mobile operators plan and deploy their networks. Initial cellular networks planned to provide wide area coverage for voice services have been very successful at establishing outdoor coverage from the beginning, but indoor coverage has required much more effort and investment from operators.

Walls, windows, and other physical barriers make it more difficult for the wireless signal to penetrate buildings, and typically a less efficient modulation scheme has to be used to connect indoor devices to outdoor macro base stations. As a result, subscribers running the same application on the same device at an indoor location on average use more network resources than if they were outdoors. For a mobile operator, this means that the effective capacity of the existing network infrastructure decreases as indoor usage grows.


With the explosive increase in data traffic, it is becoming increasingly difficult—not to mention expensive—to support data traffic generated at indoor locations with base stations located outdoors, especially if they are macro base stations mounted high on cell towers and therefore further removed from the indoor locations of subscribers that street level small cells.

Operators have started to address this issue by deploying solutions for indoor coverage, such as picocells, Wi-Fi off-load, distributed antenna systems (DAS), or femtocells. By bringing the base station equipment closer to the subscribers, these solutions greatly improve indoor capacity and coverage, and can be deployed in a focused way to transport traffic in the areas with the highest loads.


The deployment of indoor equipment has been somewhat slow to date. This is changing quickly as mobile operators are under an increased pressure to increase capacity in their networks and indoor coverage often provides the most cost-effective solution, and as an increasingly wide, cheaper and more flexible mix of solutions have come to the market. Moving from the ad-hoc solution for a few airports or stadiums to a wider indoor underlay network that carries heavy traffic loads poses significant challenges to operators, and forces them to change the way they plan and operate their networks. And this makes the transition to indoor coverage necessarily slow, and in some cases painful.

Indoor coverage requires a denser network of end-points. Operators need to manage a higher number of base stations, which increases the load and complexity on their core networks. This especially true of femtocells, which are being deployed in large numbers and reduced control from the operator on their location.

Indoor equipment is cheaper not only on a per-base-station basis, but also on a per-bit basis, and it easier and less expensive to install and maintain, more locations need to be controlled, and in many cases these locations are more difficult to protect. The biggest worry for operators is often the backhaul, unless they happen to have access to their own backhaul infrastructure. Although most buildings offer connectivity options, they may not be sufficient, not available, or overly expensive to the mobile operator. Finally, physical access to indoor locations may be difficult to secure and subject to limitations that stricter than those for outdoor cell towers. The emergence of neutral host players offers an attractive innovative business model that can address some of these issues.


Indoor coverage when carefully targeted is a very efficient way to increase network utilization—as opposed to increasing network capacity of the entire network, which may or may not be used—because data traffic is not distributed uniformly across the operators’ footprint.

In Europe, only 7% of base stations in the Vodafone network are at or near capacity. Across all markets, a few subscribers account for a disproportionately high percentage of mobile traffic. In Europe, 6% of Vodafone subscribers account for 54% of overall traffic. The thriftiest 75% of smartphone subscribers use less than 200 MB per month and account for only 17% of traffic. A similar distribution is found across most mobile operators. In the US, for instance, 5% of users account for 68% of traffic, according to Sandvine.

As a result, operators do not need to blanket their entire footprint with indoor coverage, but target it at the pressure points in their networks, which can be either public venues located within the coverage area of base stations running at capacity, or the homes of their heaviest users. The evening usage peak suggests a strong component of residential usage, which is concentrated in a relatively small number of households. Providing those with femtocells may allow mobile operators to postpone or avoid expensive network upgrades in residential areas where they typically have good coverage, but limited capacity.
Bookmark and Share

| | | | |