The man who invented ethernet is now set on establishing the ‘internet of objects’, ZigBee, linking millions of chip-enabled electronic devices across the world. Andrew Lee reports.
Bob Metcalfe is delighted to meet The Engineer. ‘I’m an engineer! I love engineering,’ beams the amiable New Yorker.
Metcalfe did, indeed, graduate in electrical engineering back in 1969. The fact that he is still keen to refer to himself as an engineer is rather endearing. Since walking out of MIT with his Bachelor’s degree Metcalfe has helped build the early internet and invented ethernet, one of the key enabling technologies of the IT revolution.
He has founded a billion-dollar corporation, published countless articles and books and been showered with engineering honours. He even has a law named after him, Metcalfe’s Law: the usefulness or utility of a network equals the square of the number of its users.
Metcalfe is in Cambridge to spread the word about ZigBee, the wireless networking standard with big ambitions to link billions of devices. Variously known as the ‘wireless mesh’ or the ‘internet of objects’, a ZigBee network would see low-cost, low-power radio-enabled chips linking almost every conceivable object to the wider world.
Cheaper and less power hungry than the better-known Bluetooth, ZigBee – technically known as 802.15.4 – would have a host of everyday applications. Metcalfe gives an example: after a power cut many of us have been faced with the timer display on our microwave flashing 8888.
The oven’s microprocessor, adept at controlling the cooking time of frozen chicken, is essentially dumb. Cut off from the outside world, it has nobody to tell it what time it is except its human owner, who is condemned to a 10-minute hunt for the instruction manual and a bleep-infested battle with the controls.
As part of a kitchen-wide ZigBee network, however, our stupid microwave would merely ask its control unit for the correct time and reset itself.
This mundane example gives a flavour of how ZigBee works, but fails to encompass the massive scale of deployment envisaged by Metcalfe and its other proponents. Industrial automation is an early target market, the countless domestic electronic devices used in homes around the world another.
‘The potential market is huge,’ said Metcalfe. ‘Less than two per cent of the eight billion microprocessors shipped each year are networked. Then there are the people who are not even using microprocessors because there hasn’t been networking available to date, and therefore not enough value to put a microprocessor into their products.’
Metcalfe’s support for ZigBee is not without self-interest. Polaris Ventures, the venture capital group of which he is a partner, is a significant financial backer of Ember Corporation, a US company at the forefront of wireless networking technology development.
Ember recently strengthened its position in the fledgling ZigBee sector when it bought the portfolio of 802.15.4 integrated circuit technology, developed by Cambridge Consultants, and hired the engineering team behind it. Ember will work with its newly acquired UK interests to develop single-chip ZigBee systems that can be deployed in new products.
Metcalfe and Polaris have put their money where their mouths are as far as ZigBee is concerned. Where networking technologies are concerned, it would be a brave soul who bet against Metcalfe’s judgment. In 1973 the American invented ethernet, the now ubiquitous local area networking standard used to connect billions of computers and their peripherals.
Metcalfe was working as a researcher at the Xerox Palo Alto Research Centre when he made the breakthrough that would change his life and shape the development of the global IT sector to this day. ‘I was given a problem that nobody in history had ever faced before. How do you network a building full of personal computers? I made a few good engineering decisions, well prepared by my education and environment, and out popped ethernet,’ said Metcalfe.
The American shares four patents on ethernet, a technology that shows no sign of running out of steam. Some 200 million ethernet ports will be shipped in 2004, a figure Metcalfe admitted was inconceivable in 1973 despite the pride he and his colleagues in California took in solving their networking problem so effectively.
‘There was no grand vision. It’s true that one of the things they teach you at engineering graduate school is good taste in engineering. There was a certain high-mindedness about it, but nothing that would smack of 200 million units a year by 2004.’
Metcalfe believes the new breed of networking technologies, and ZigBee in particular, are set for even more spectacular growth. ‘The excitement for me is drawing parallels with ethernet’s take-off, which started 31 years ago and ZigBee and Ember, which are all taking off in a similar pattern.
‘I remember doing that, I remember that happening and part of my delight in the project is reminiscing. In fact, I suspect Ember and ZigBee are going to take off much faster,’ he said. ‘People know how to build chips now, and we didn’t in those days. Everyone has seen the ethernet model and is much more reassured about the emergence of a standard. Plus the market is so much bigger. I don’t think it’s going to take ZigBee 31 years to get to 200 million units a year.’
Metcalfe’s point about the emergence of a standard is crucial to why he and ZigBee’s other evangelists are putting their faith, and their cash, behind the technology.
ZigBee is by no means alone in spotting the potential for low-cost, low-power networking for both industrial and consumer applications. Proprietary systems for specific products abound, but ZigBee-based developers such as Ember are betting that users will quickly cotton on to the benefits of a technology that operates across industry boundaries.
The ZigBee Alliance, the international body promoting the technology, boasts support from the likes of Honeywell, Philips, Samsung, Motorola and the UK’s Invensys.
‘We know what can happen to a network standard when it takes off because we’ve seen it with ethernet,’ said Metcalfe.
His new role as a venture capitalist is merely the latest career departure for Metcalfe, who shows absolutely no inclination to slow down as he approaches 60.
Before inventing ethernet, he had already worked on the very earliest version of the internet. 3Com, the company he set up in 1979 to exploit the commercial potential of ethernet, turned into a billion-dollar corporation. Metcalfe spent most of the 1990s as a magazine publisher, author, newspaper columnist and all-purpose technology pundit. He also managed to fit in academic posts at, among others, Stanford and Cambridge.
In fact, Metcalfe’s return to Cambridge causes him to ponder that old chestnut: how does the UK stack up against the US when it comesto commercialising hi-tech innovations?
‘I spent a year at Cambridge a decade or so ago. The question at that time was why Silicon Valley had start-ups that scaled up – Apple, Intel, 3Com and so on – and Cambridge didn’t.’
Metcalfe believes he knows the answer. ‘It isn’t because English people are stupid, it isn’t because there’s no money in the UK to invest in start-ups, it isn’t because English people can’t market products. In Cambridge at that time, business was not a proper occupation. When I was here management studies was new that year – the first MBAs in 700 years. They were having trouble getting management professors into colleges because business was not seen as a proper occupation.’
Metcalfe senses that things have changed. ‘Those management professors have had a chance to be around for 12 years, and I think that’s probably helped. I drive around here now and the buildings seem bigger. There’s some scale occurring here.’
As for Metcalfe himself, he has a 10-year plan. ‘I want to be the best venture capitalist in the world. This is not easily done,’ he added with dry understatement. ‘But I’ve started my apprenticeship.’