What bandwidth should operators allocate for a typical urban small cell? Too little, and they may throttle the capabilities that 3G and LTE can provide both now and in the future. Too much, and it could impact the cost significantly.
Several small cell backhaul vendors have told me in the past that 100 Mbps is perfectly adequate for individual urban small cells, even including the latest multi-mode 3G/4G/Wi-Fi products. While macrocell backhaul links are increasing to 1Gbps or more, this has been seen as overkill for individual on-street units. But is this the right choice when equipment lifetimes are budgeted for 5 years or more? A few backhaul vendors are offering higher speed products that may satisfy both bandwidth and budget.
Of course fibre would be the first choice
The high speed and reliability of a hard-wired fibre connection would make this the first choice from a purely technical standpoint. Fibre is implicitly capable of 1Gbps speeds or more, and throughput is limited by commercial pricing rather than technology.
But fibre takes time to install. Many urban small cells will be mounted on street furniture such as light poles, bus shelters and fixtures which are not hard-wired today. AT&T had to make substantial changes to their back-office systems to enable it to handle orders for wired backhaul to be placed for individual street furniture rather than just to a building street address or postcode. I doubt that all operators would have the ability to do that yet.
A combination of wired and wireless would be the most flexible
Inter-site distances in urban canyons are said to be about 160 metres according to one US Tier 1 operator. A daisy-chain of street furniture mounted small cells could easily be connected to a fibre-enabled hub on a rooftop or streetbox, often co-located with an existing macrocell.
Small Cell network planning is still a bit of a "black art" when selecting the best sites from those available and some planners might prefer to trial a few locations to assess traffic usage patterns before committing to one. The smaller coverage footprint of an urban small cell means that locating it 10 or 20 meters away from traffic hotspots could reduce its impact significantly. Wireless backhaul provides the flexibility to deploy quickly without committing to a long lifetime lease of a wired connection.
The choice between longer lifetime and lower backhaul equipment cost
Backhaul should be future-proof. The total cost of ownership throughout its lifetime is less about the initial cost of the equipment and more about durability, long lifetime and fewer site visits. This may lead some operators to favour equipment capable of higher speeds than they need from the outset, knowing that the equipment will have a longer lifetime.
Backhaul topologies where several small cells are daisy-chained (e.g. down a street canyon) may require higher throughput to handle the aggregate capacity from a number of cells. The traffic from each site could peak when a combination of 3G, LTE and Wi-Fi are all used to their maximum extent.
A plethora of choice
Many vendors offer a range of technologies and products to address this market, all at various stages of maturity, performance and price. Many have roadmaps towards 1Gbps or more and several offer these higher speeds. Network operators have to assess and make the tradeoffs between technical performance and price.
It will only be when the industry starts placing some large orders that we'll see the price point driven down. I've heard of a target of $1000 per 100Mbps link indicated by operators, but believe we are still some way off this until volumes increase dramatically.
Operators conduct extensive trials and testing programs
Several tier 1 operators are actively conducting their own tests and trial programs, assessing the various technologies and options available. I've seen several present their results at conferences, such as EE (in the UK) and Sprint (in the US). The process is iterative with operators feeding back their results, suggestions and recommendations to vendors who can then revise and improve their product.
These tests aren't simply about throughput or performance, but check everything from operation in different weather conditions, how links are affected by the twist and sway of lightpoles they may be mounted on to the overall time/complexity required to install and maintain the equipment.
The tests are inclusive across all types of backhaul technology including new and innovative ones.
Free Space Optical proves it can achieve gigabit speeds
One example is a technology we looked at last year - Free Space Optics (FSO), which uses low power infrared light beams to carry 100Mbps or more across distances of a few hundred metres to a few kilometers. Key benefits include licence free spectrum (surely the use of light is unlikely to be taxed), low equipment cost, freedom from RF interference and high performance. Downsides are the need to align both ends of the link during commissioning and ensure a continuous clear line of sight, not unlike 60 – 80 GHz radios. It also has to overcome a reputation of potential disruption in bad weather (eg fog, heavy rain) which it no longer deserves.
Since then, one FSO vendor – Polewall – has further developed their product, the StreetHopper, using a 1.15 million Euro EU research grant to build a multipoint link that delivers reliable and consistent 1Gbps data rates. One US Tier 1 operator conducted extensive testing of the product as part of their wider small cell backhaul technology testing program, reporting very low, latency, jitter and packet loss and overall extremely positive results.
Polewall are very upbeat about the possibilities the technology opens up and will demonstrate a 10 Gbps link within a few months. The company competes most directly with 60GHz short range/lightly licenced wireless links that are now being commercially deployed for short range wireless broadband links. They remain confident that they can maintain a lower price point because of the simpler components used – satisfying the conflict between high speed, low cost and long lifetime.
Summary
The peak backhaul bandwidth for an urban small cell is likely to increase over time, especially when multi-mode 3G/LTE/Wi-Fi services are used concurrently.
Operators will need to make some tough choices about the trade-off between peak speeds, equipment cost and planned equipment lifetime.
Selecting backhaul products capable of higher speed is one option that could extend the payback period for capital investment.
Several backhaul products are available today that offer those speeds, including some new innovative options such as FSO, 60GHz and others.
