With data traffic growing quickly but revenues remaining static, the cost of carrying each bit of data across a wireless network needs to drop substantially. A recent report by Analysis Mason suggests that a 50% cost reduction per bit is required by 2015 and includes small cells (specifically Wi-Fi) as a critical factor to achieve that goal.
While several studies have shown that small cells are substantially cheaper than macrocells, one aspect worth highlighting is their comparative utilisation levels. Small cells may be idle or handling minimal traffic for longer periods of the day. This means that the small cells must be disproportionately lower cost per unit than their macrocell cousins – a factor that is quite achievable but may be overlooked.
Why is the last mile often more expensive than the long haul
This concept can perhaps be better explained through an analogy with public transport. For example, flying longer distances by airplane has become remarkably low cost. Planes are almost always full in my experience and their utilisation kept very high. Conversely, when arriving back at the airport it is not unusual to be the sole driver of your car which has sat unused for days. In some cases, the cost of the car trip home can exceed the price of the air ticket.
Utilities such as gas, electricity or water are also provided across wide areas with major pipelines carrying the same service shared by thousands of customers. It is again that "last mile" of connection that attracts the highest proportion of the cost, with each individual pipe attracting maintenance costs per mile that are disproportionately high.
Small cells can offer high peak capacity
A small cell may have a significant amount of capacity to offer. Many of today's 3G femtocells can handle peak data rates of 14Mbit/s, matching those of the latest iPhones and potentially track a hundred or more users on standby. By comparison a macrocell may be fitted with multiple carriers (2x or 3x the frequency allocation) and multiple sectors (3x or more). Other advanced features such as MIMO may add further gain, reaching perhaps 10x that of a typical single small cell.
However the small cell is unlikely to get the same level of utilisation as a wide area macrocell. The small cell may be targetted to handle shorter peak periods of high demand with longer periods of minimal use. The macrocell may be covering a mix of different types of buildings and use, including a mix of residential/business/retail where a small cell is more likely to be targeting a single behaviour.
Residential femtocells benefit from free site rental, power and backhaul; metrocells don't
Once the one-off capital cost of a residential femtocell has been paid – typically in the $100-$200 range – then the operator does not have to pay ongoing costs for site rental, power or backhaul. These are all covered by the subscriber themselves. It allows the network operator to scale up to many hundreds of thousands of femtocells without attracting a high ongoing OPEX cost.
This isn't quite the same for public area/outdoor metrocells, where some contribution may be required to cover the site rental, power and backhaul elements.
- Site rental might be a small fee paid to the town council (or similar owner of the public areas affected). I would expect this to be done through a general agreement, with dramatically lower rates than for large tower sites we are familiar with today.
- Power requirements for small cells are comparatively low, say in the range 20-200Watts depending on traffic levels (so perhaps less than $100 per annum), higher for rural metrocells with a longer range. Again this is proportionately much lower cost than a macrocell which may consume several kilowatts.
- Backhaul is likely to be a mix of wireless and wireline (fiber). The peak capacity of a set of small cells can be quite substantial (say 1Gbit or more for the latest LTE small cells)
In addition, the cost of public access small cells tends to be higher because these devices are:
- Toughened to handle operation in outdoor, unsupervised areas
- Often have higher traffic capacity
- May have higher power levels to extend coverage area
- May have more complex external antenna arrangements
Implications for network planners
Analysts forecast large numbers of small cell deployments in the years ahead, to meet burgeoning demand for wireless data. Figure of anything up to ten times the number of cellsites are suggested in high traffic areas. The decisions of where, when and how quickly to roll out these smaller metrocells are likely to be made using a different set of criteria than for macrocells. The economics of small cell deployment are significantly different than for larger sites.
The lower utilisation levels, more apparent as the number of small cells increases radically, mean that this lower proportion of traffic carried must be financially justified. The costs of deploying and maintaining an individual small cellsite must be at least ten times less than a macrocell – something that should be quite achievable.
As the industry moves focus from residential towards public area metro and enterprise femtocells, the additional costs both of the small cells and the associated operating environment need to be recognised.
Careful cost/benefit planning will be required to ensure that the financial benefits of each new small cell deployment exceed the total CAPEX and OPEX costs of deployment. This will lead to different behaviour and processes within the network operator's planning departments, to ensure that the right decisions are made each time.