Comparing the cost per bit of macrocells and small cells

A major business driver for adopting small cells is to address expected capacity constraints in mobile networks to meet rapidly rising traffic demand. ThinkSmallCell recently led a panel session at the Basestation Conference earlier this month, delving into the key factors that make the business case for small cells stack up. We've outlined some of the themes discussed and listed the key factors affecting investment choices.

Megabytes vs Megabits

A common metric to calculate how much data is passing through our networks is simply to total up the quantity. We read of how many Petabytes (million Gigabytes) of data are consumed daily, and that's what end users are billed for.

But perhaps a better metric is the bandwidth, in Megabits per second, that we receive. That's what customers are measuring and determining their performance from. It's also being used to promote adoption of premium 4G services offering higher speeds.

Where large numbers of users congregate in urban areas, the total bandwidth required in a given area is high. Operators must deliver this within the constraints of the spectrum they control, measuring capacity density in units of bits per second per Hertz per square metre.

What's the end-user value of a Megabyte of wireless data, and where does it vary?

As consumers, we take ubiquitous access to wireless data almost for granted these days, getting annoyed when it's not available or too slow. Breaking out the specific technical characteristics that end-users relate to, we can identify:

• Coverage: We expect the service to work wherever we go, from remote and sparsely populated rural areas through to the deepest underground tunnels in our metro systems.
• Mobility: Being able to use the service when moving around, even quickly when in a car or train, and seamlessly maintain the connection as we walk in and out of buildings. Nomadic services, which work only in specific places when we stop and sit around, are useful but can have less value.
• Data rate: Not peak theoretical speed, but the real-world speeds we experience. While peak speeds are faster using HSPA+ and multi-carrier LTE, these might be throttled by overloaded cell sectors, backhaul throughput and even core network bottlenecks.
• Latency: Highly interactive applications and websites with many images appear to work much faster with low end-to-end latency. LTE has a considerable advantage here, with shorter response times designed in. This can be perceived by end users simply as higher data rate, but is a different characteristic.
• Jitter: For interactive services, such as voice, a wide variation in packet delay results in lost packets and garbling/dropouts in the audio stream. As we migrate voice and video-calling services across from circuit switched to packet switched, substantial engineering is required to support these services at a similar or better quality level.
• Reliability: Telecoms networks are designed to be "always on" with 99.999% availability. We have come to expect our smartphones always to work and are surprised when the service is offline.
• Roaming: For foreign travellers, the ease and simplicity of accessing service in the same way, albeit at a premium price.

Capacity density itself isn't a feature that end users would notice directly, although it affects their data rate in congested areas at busy times.

Charging a premium price for data

Today's mobile networks don't discriminate in pricing between a Megabyte of data delivered from a small cell or a macrocell. Most do charge per Gigabyte, with a few brave networks still offering unlimited data plans.

Several networks have successfully priced their LTE services at a premium. Most will capture increased revenues from LTE because their users consume more data, hence buy larger monthly data allowances.

Public Wi-Fi data use is often included for free or at minimal cost because such services are nomadic (i.e. don't include mobility, have limited coverage areas and variable levels of speed/latency/throughput).

Small cells can increase the value perceived by end users. It's not just the extra capacity they deliver, it's because they also increase the end-user data rates achieved, reducing both latency and jitter. Unlike Wi-Fi alone, they can deliver a more consistent service that includes full mobility that seamlessly hands over as you walk in and out of hotspot areas.

Expanding the existing Macrocell installations with extra capacity

Network operators will naturally look at making the most of their existing macrocell network first. Incremental expansion of installed sites will be their first choice. Options include:

• Sector splitting, adding antenna and basestation equipment which doubles capacity. A typical three sector site can be expanded up to six sectors, effectively doubling total capacity. Sector splitting also includes placing remote radio heads slightly further away from the main site, connected by dark fibre or special dedicated high speed wireless links.
• Adding extra spectrum: This precious asset enables extra capacity by allocating more frequency bands per sector, again needing additional basestation equipment onsite.
• Refarming spectrum from 2G to 3G to 4G: Later generations of radio technology are more efficient and deliver higher data rates. This requires basestation equipment upgrades and more end users to have the latest smartphones to make use of the newest technology.

All of these options involve additional equipment on site. Some involve extra or replacement antenna, which can affect planning/zoning permits. Backhaul capacity to the site is also likely to need enhancement.

A benefit of this macrocell expansion approach is that each macrocell covers a larger footprint than a small cell. The expanded capacity can serve a larger area and a wider mix of users, handling different use cases in slightly different places throughout the day. This could include both indoor and outdoor uses.

During the initial rollout of LTE coverage, with its associated extra new spectrum, it makes good sense to make best use of the existing sites and infrastructure. Once capacity limits take effect, the options to expand capacity density by using small cells become more attractive.

Expanding network capacity through small cells

Practical experience from network planners today report that the hottest parts of the network (those with the greatest traffic densities) are often also growing at the fastest rates. There are technical limits of how much expansion is possible at existing macrocell sites and at what point the cost of extra macrocell equipment remains the cheapest option.

Outdoor small cells, often called Metrocells, come with a number of additional costs. It's not just the physical hardware box itself, but also the backhaul, provision of power, installation (including planning permission) and ongoing management.

The cost per Megabyte of macrocell capacity expansion increases after the early "quick wins" reaching a tipping point where it becomes more economical to install a small cell layer. With traffic demand growing quickly, operators will need to be in a position to deploy them as and when needed. Those conducting trials and vendor selection now will be in a better position to adopt them when the time is right.

Indoor small cell capacity

The case for indoor small cells is often easier to make. For a start, costs are lower. The products can be designed for a more benign environment where they don't have to be weatherproof, tamperproof and resilient to environmental fluctuations. Electrical power (which can be delivered via Power over Ethernet) is often free, wireline backhaul usually easily accessible and building owners often co-operative and even welcoming.

Offloading data traffic from an outdoor macrocell can be even more valuable than appears at first sight. The proportion of resources which a macrocell has to allocate to penetrate in-building will be much higher than that to serve a nearby outdoor subscriber. The Megabyte of data delivered at higher speed to the indoor customer will also be perceived to be of higher value.

Factors affecting the trade-off calculation

While the mobile network industry is at a stage of deploying the initial LTE service nationwide, it will naturally focus on building on its existing cellsite base.

We are already seeing several countries with networks further along the path of LTE adoption moving to the next stage of capacity expansion. This involves assessing the cost of squeezing even more out of their busiest cellsites after the easy options have been exhausted. Whether that is best achieved by deploying a number of small cells or using additional remote radio heads and/or more sophisticated LTE-Advanced features will depend on individual circumstances.

Small cells are the more cost effective option once the installed macrocell sites have been expanded to their normal limits.

Operators in different countries are constrained by many additional factors, such as:

• Limited spectrum and/or spectrum at less useful frequencies
• RF power constraints (e.g. Italy, which has tight limits on total radiated power from any installation)
• Zoning/planning limits for additional antenna, including constraints set by landlords
• Existing coverage holes, especially in-building
• Large installed base of repeaters, which don't add capacity and may now be adversely affecting performance of neighbouring areas
• The need to adopt Voice over LTE, which requires higher network performance (especially packet delay variation) than a data-only service
• Their approach to Wi-Fi as an offload/alternative service delivery option

Summary

The cost of delivering a Megabyte of data traffic using a small cell should be cheaper than via a macrocell, especially indoors where costs of power, site rental and supervision may be incidental.

In high traffic areas, where a small cell can be targeted to offload substantial amounts of data from the macrocells, the business case to do so is easy to justify.

Where quality of experience matters, using small cells indoors to deliver a higher value service – such as voice – is also easy to justify and cheaper than other methods. This is particularly important for networks with less spectrum and/or less in-building coverage.

As we watch traffic levels continue to grow, and the network migrate to wider use of LTE, it seems likely that the cost effectiveness of small cells will make their mark. The vast majority of network operators already have them in their roadmap and will adopt them at that crossover point when they become cheaper to deliver capacity than other methods.

Do you agree/disagree with the analysis above? Why not comment below (anonymously if you prefer) with your own views.