• silicon.com
• Technology
• Networks

By Peter Cochrane, 7 September 2006 11:15

COMMENT

Written whilst flying Gwangju to Seoul on KE1302 and despatched to silicon.com via one of several free wi-fi connections at Seoul Airport

As a post-graduate student in the 1970s, I attended a broad selection of courses beyond the normal engineering and science spectrum on the basis that I would most likely have to manage people, money and projects at some time in my life. Among these courses was one on the basics of accounting and economics. Here I learned about the old practices of double-entry book keeping, discounted cash flow, net present value and - most valuable of all - whole life costing.

It wasn't long after this, when I had finished my PhD and had moved into systems design and engineering, that I found myself deep in the debate on the deployment of the then-rapidly maturing technology of optical fibre. As the loss of glass fibre per kilometre dropped and availability of lasers and pin diodes advanced, the economic wheels turned in favour of replacing all the old copper trunk (long distance) lines with fibre. And so one afternoon over coffee, the decision to switch to optical fibre was made in the UK, and similarly in the US and Japan.

Around seven to 10 years later optical fibre was everywhere in the UK. All the copper and microwave radio long-lines circuits had been relegated to the files of history, turned over to a lesser role of broadcast distribution and emergency stand-by use. The original economic calculations and estimates were spot on! Massive savings were realised in equipment and people, network performance was improved, operating costs fell and customers benefited enormously.

This pattern of events was echoed across the planet as country after country came to the same conclusion and opted for fibre. Furthermore, undersea cables containing optical fibre were laid across the Atlantic Ocean starting in 1986, to be followed by the Pacific and all the seas, to the point where satellite communication was relegated to the insignificant and became largely irrelevant for major trunk applications.

There was a very obvious next step - to advance fibre into the local access network and get right down to the home and office. At first the technology and economics looked promising, and in 1986 the first optical fibre trial system to undercut the cost of copper was installed in the UK. But then after extensive system development worldwide, the economics of the mad house kicked in globally. It was as if people forgot all they had learned, and the emphasis flipped to a simple-minded upfront costing model. It was as if the whole industry lost the plot overnight!

To be fair the guardians of the copper access network had seen their long-line brethren retire early or find new jobs as their numbers shrank by 90 per cent with the advance of fibre, and naturally enough, the promise of DSL (digital subscriber line) technology - which runs over copper wires - seemed to significantly discount the need for fibre in the short and long term.

But as it turned out, DSL only delivered less than 10 per cent of what was originally promised in the early 1990s, and the customer appetite for bandwidth grew 1,000 per cent faster than the forecasts! All that was predicted by many in the industry but largely discounted because it was economically and politically unpopular. The planet had decided that copper was it and no amount of proof or reason was going to change the decisions made and directions already selected.

It was all a bit of a bummer and the decision to stay with copper in the access network has, to date, cost most nations an arm and a leg. But what is particularly damaging is the apparent and continued inability to make any intermediate correction of what was a really bad decision in the first place.

Worse, it seems as if all economic decisions are now exclusively made on the basis of upfront cost. What really fascinates me is that the people inside a company who are taking such a simple-minded view wax lyrical about the quality of their latest badged and exclusively designed goods. They take a lifetime view at home and a 'cheapest is best' approach to everything for their company!

This past week I have been at my first optical technology conference in a decade and the biggest non-issue being extensively discussed was still 'fibre to the home' (FTTH). But the advances in optical fibre, lasers, devices and microprocessors over the past 20-plus years have rendered all the arguments for FTTH absolutely unassailable. However, the cheapest-is-best merchants persist.

So here is the route to the biggest transformation in telecom networks ever:

Water ingress in copper cables typically accounts for around 50 per cent of all the faults found in the access network. The cost of repair and keeping cables dry is enormous in materials, manpower and loss of revenues. Fibre on the other hand can be wet - glass is impervious to water.

People employed in the switch and hub sites of the access network to reroute copper pairs, and repair all manner of faults, do in the course of their work inflict around 25 per cent of all line faults. With fibre, rerouting can be remotely programmed and human intervention at a physical level is not required and saves around 95 per cent of all human activity in the networking aspects.

Fibre reach is at least 10 times that of copper and therefore removes huge amounts of electronics from the field as well as reducing the number of switch and hub sites. For example, in a copper network with 1,000 switch sites, the move to optical fibre would see a reduction to only 10 or so. The continual saving on unwanted real estate and the upfront - one-off sale - recoups vast amounts of money.

Energy consumption is vastly reduced with fibre access, along with huge reductions in the need for battery back-up, standby generators and access points.

Operational and business support systems can be dramatically reduced in scale as the amount of plant to be managed with fibre compared to copper is reduced by around 100-fold depending on the network geography, topology, operations and services delivered.

People requirements for a fibre network are around 10 to 15 per cent of the copper equivalent.

Path transparency of fibre networks gives a very effective future-proofing against new technologies, services and customer demand. In short, fibre gives us the ultimate and near infinite bit-transport pipe. Unlike copper, it does not suffer from bandwidth restrictions, harsh attenuation and cross-talk limits at high frequencies, and is indifferent to signal format.

Operational Expenditure (Opex) falls year-on-year with technology advances and new, unlimited service offerings.

Future Technology coming down the optical and micro technology pike is going to boost all of the above arguments even further over the next decade, and I can see the prospect of network people requirements falling to five per cent of today.

Well, that's most of the argument from 1986, and it is even more applicable today. It really is that simple, and moreover it is now even easier to do now than before. Just start with new install, stop buying more copper and get fibre rolling. The savings will start to bootstrap the overall investment for network transformation.

I am not a religious man but if I were I would be tempted to think that god is a communications engineer as he gave us abundant supplies of silica (for glass) and silicon (for chips). Both are inert and safe, easy to manipulate and work with and have ideal electrical and optical characteristics for computing and telecommunications. Unfortunately, mankind discovered copper first and went off on a very expensive detour!

I jest of course, for without copper technologies we would never have discovered the magic of silica and silicon. But now the time of copper is long gone and we should gracefully bid farewell!

Imagine what would have happened if our Victorian forefathers had adopted the simple-minded, soda straw, 'one-at-a-time issue', economic principals of today. There would be no telecommunications network. The initial rollout cost would never have been justified!

What has happened? Might it be that this is all a manifestation of an increasingly limited or failing education system globally? Are we really becoming so simple-minded that we can't simultaneously appreciate and deal with several broad arguments spanning engineering, science, business, economics, social and political consequences as well as competition?

I think I'll leave wireless in the access network out of the picture for now and address it in another blog!