The relatively recent, but welcome, public recognition that broadband telecommunications will have a significant impact on our economic future, has led to much discussion and debate about how we as a nation should develop this broadband future.
As a technologist I have found that much of the discussion has been somewhat equivalent to saying “if we could defy the laws of gravity then we could...”, so in the interest of informed discussion I have decided to write a series of blogs which I am entitling “The Technology Laws of Telecommunications Economics”.
While most of the debate has been conducted around the economics of the choices, there has been little recognition that the reality of the technology trends does impose stringent constraints on what is possible and desirable.
Law 1: “Moore’s Law is the source of much goodness, but does not solve all problems”
Over 40 years ago, Gordon Moore predicted that the number of transistors per chip would continue to double approximately every 18 months. This is an astonishing prediction – it is saying that every 15 years or so, the capacity of a chip will increase 1,000 fold. Another way of looking at this is to say that technology that cost $100,000 in 1992, is now available for $100. Wifi base stations are an excellent example of this.
The reason Moore’s Law is the “source of much goodness” for telecommunications is that it creates an unprecedented ability to process signals. All telecommunications systems have the same characteristics: you put a signal in at one end, that signal is transformed by the telecommunications channel and infected with noise, so at the far end you need to remove the noise and transform the signal back to the original. These latter 2 operations involve computing (actually signal processing) and the more computing you can do the better you are able to remove the noise and complete the transformation back to the original signal. Moore’s Law says that in 18 months time I shall be able to do twice as much processing of the signal than I can do today.
From this simple fact follows all the improvements in telecommunications systems that we have witnessed over the last 30 years. For example, a lesser known law, “Cooper’s Law”, states that the capacity of a radio system measured in bits/Hz/unit area doubles approximately every 2.5 years. This has proven to be roughly true since about 1900 soon after Marconi invented the radio. The reason why Cooper’s Law is valid is that Moore’s Law enables ever cleverer and more sophisticated operations to be performed on the radio signal. Similar, Neilsen’s Law, which states “A high–end user’s Internet speed doubles every 20 months” also follows from Moore’s Law. DSL technologies are a creature of cleverer and cleverer computing.
Computing operations are not new. There is no reason why Charles Babbage could not have set Ada, Countess of Lovelace (the first computer program) to work on the task of creating the algorithms for DSL in the 1830’s. However, it would not have made sense unless there was an economic way of implementing these algorithms, which is of course where Moore’s Law comes in. So what Law 1 implies is that while many things are possible, they do not eventuate unless their utility exceeds their affordability.
Unfortunately, not all aspects of telecommunications are subject to the benefits of Moore’s Law. Labour (and associated costs), copper wires, land for towers, ducts, cabinets, in fact everything that is not amenable to electronic implementation, all go up in cost, some like copper and fuel by a massive amount. If these costs were a negligible portion of the overall cost (as they are in the computing industry), then overall costs would drop with the rate of Moore’s Law. Unfortunately for telecommunications these ‘non-electronic’ costs represent a significant fixed cost that constitute a large portion of the total cost of a telecommunications system. This leaves telecommunications in a somewhat invidious position because some aspects are changing dramatically fast and others plod along very slowly. Thus, many of the ‘popular’ (or should I say ‘populist’) analogies with other industries (e.g. computing, utilities) are facile and meaningless. I shall take this up in subsequent blogs describing other ‘laws’.