Reliable, high-speed access to the internet is vital for many businesses and is becoming very important for home users too. Everyone is interested in improving peak speed and throughput, so it is useful to think through the way network bandwidth might develop.

Something akin to Moore’s Law – which suggests the power of processors will roughly double every 18 months – also applies to bandwidth. But the technologies for bandwidth tend to be quickly superseded by new ones.

It’s pretty much a given that the core network can provide as much bandwidth as required because of the huge amount of unlit fibre already in the ground – the “glut of glass” as it was known. As a result, the problems largely boil down to making connections from the customer’s site into the core public network.

Today, there are three ways of connecting: copper, fibre and radio. Copper is reaching its limit with VDSL and ADSL2+. It’s not going to be possible to get much more than 30Mbit/s, and that only if you’re close to the telephone exchange. It won’t be long before copper fails to provide an acceptable service for a significant number of users.

Fibre is, of course, the answer to all our prayers and it will be a long time before a fibre connection is unable to cope with the needs of subscribers. However, not every location has fibre, and it would be a massive task to bring fibre to all residences.

Business locations fare rather better and they tend to have fibre connections available. Fibre may also help to extend the life of copper if DSL equipment fed by fibre is installed in street cabinets.

Radio is the dark horse of the access network, and while it is unlikely to dominate, there will be several gaps that it can fill. Obviously, radio has a monopoly on mobile access. Current developments will eventually lead to rates of up to about 20Mbit/s.

Fixed radio access can offer similar rates to DSL today and can reach further. I think we will also see a niche for radio when users start to demand high speeds but are too far away for DSL and have no access to fibre.

In theory, radio systems could be developed to offer rates similar to fibre and provide service to areas without fibre. However, this niche will only be filled if fibre rollout is slow and if enough locations for transmitter sites can be found.

Another niche for radio will be to provide backup for fibre connections. I think this will become more important as companies realise that all their traffic is often carried by just one fibre. So radio could provide companies with extra resilience by adding an alternative route.

Traditionally, new operators have used radio to compete with the incumbents who have landlines, but I think radio will also be used more extensively in future by the incumbents to fill their gaps and provide a full service. We will even see unbundling of radio access connections, bringing more competition.