HetNets – Heterogeneous Networks – is a relatively new term that crops up more frequently in mobile networks. It refers to using a combination of large and small cells (macro/pico/femto) with different radio technologies (2G/3G/LTE, even Wi-Fi), all working together to provide the best coverage and capacity possible. In other words, combining all the options for technical fire power of all shapes and sizes to meet the burgeoning demand for wireless data.
Ericsson promoting HetNet concept strongly
Ericsson has been particularly keen to market this term. They rarely use the word femtocell, having dismissed the business case for residential femtos a long time ago. However they do recognise the need for small cells, particularly in public outdoor and urban areas. This Ericsson slidedeck from March 2011 succinctly explains the point. This Ericsson technical paper goes into more detail, outlining three stages of network capacity expansion:
- Improving the macro layer: i.e. adding more antenna, using more spectrum (2nd/3rd carriers) at existing cellsites
- Densifying the macro layer: i.e. installing more macro cellsites
- Complementing the macro layer: i.e. installing an underlay of pico/metro-femtocells and using Wi-Fi
I expect we could see many network operators follow this lifecycle, and have observed different adoption rates depending on what options are available and competitive pressures/pricing. It is relatively easy for existing network planners to adapt to this way of working, which simply extends their current design and planning practices – albeit scaling up the number of cellsites and basestations they need to manage. The configuration of each cell becomes more complex, with a growing interdependence between the settings (eg power levels, frequencies etc.) used within each layer.
Unco-ordinated versus Co-ordinated Heterogeneous Networks
Ericsson make a bold statement that substantial extra capacity improvements are achievable if the macro and metro-femto basestations are fully co-ordinated, quoting a capacity improvement of 3.2 compared with an unco-ordinated system of macro and ad-hoc metro-femto. This could be read as an attempt to justify operators buying all of the network from the same supplier, where extra proprietary optimisation features can be used to improve the system performance.
Nonetheless, co-ordination between femtocells themselves has been available as a proven solution for some time. Ubiquisys have deployed their "self organising grid femtocell" capabilities already, where each femtocell communicates with its neighbours to directly manage traffic levels and automatically adjust the range/power levels to balance the load.
Future standards, especially for LTE, include SON – Self Optimising Network – which automatically tune and adjust the basestation parameters to improve performance. These are designed to operate in multi-vendor environments, such as where the macrocells and picocells are supplied by different companies.
What network planners want to hear
The three stage expansion approach which Ericsson outline does seem to match what network planners want to hear today. Many (but not all) already recognise that simply expanding macrocells and installing a few extra ones won't keep up with today's rapidly expanding capacity demands. A new "underlay" of small cells – whether you call them metro-femto, picocell or something else – is essential to provide the additional capacity and high data rates that customers expect.
How closely these cells need to interact for optimal performance remains to be seen. By introducing the term HetNet, there is acceptance that a variety of small cells complementing the existing macro network is required. Ericsson are taking this a step further and arguing that close real-time interaction is needed between all cellsites in order to achieve optimal performance.
Metro-Femto demand being matched by vendor investments
Equipment vendors are working hard to meet this market niche. This has fed through to chipset vendors who have made a spate of product announcements recently for product suitable to build metro-femto products – Freescale, Texas Instruments, Picochip, Qualcomm have all been visible recently. This has fed through to vendors, both familiar and new entrants, developing a range of small cells for both LTE and 3G (including combined units).
For the end consumer, this should mean much improved data speeds and capacity in the urban areas and other high traffic hotspots in the long term.
There can be no doubt that the industry is adopting the femtocell approach for outdoor/public areas, whether it chooses to call it small cell or HetNet or anything else. Major investments are being made. Even Ericsson, who have been sceptical about residential femtocells, are actively promoting the concept today – suggesting that metro femto is rapidly becoming a mainstream solution.
In addition to the basic product hardware, self-configuration and optimisation software features are likely to become much more important to network operators because they will achieve much higher return on investment and better service to their customers.