Pauline Rigby: April 2008 Archives

I've just read an excellent article in IEEE Spectrum magazine on how the pace of innovation in electronics brings its fair share of problems as well as solutions.

Technology obsolescence is a problem that affects everyone — every time I turn on my 4-year-old home PC, I think about buying a new and faster one. But in some industries, especially those where the technology is expected to last a long time, or be ultra reliable, the problem is particularly acute. Telecommunications provides the author with some fine examples.

"Consider one major telecommunications company (which wishes to remain unnamed for competitive reasons) that typically buys enough parts to fulfill its anticipated lifetime needs every time a component becomes obsolete. Currently, the company holds an inventory of more than $100 million in obsolete electronics, some of which will not be used for a decade, if ever.

That sounds expensive! What's more, the issue of obsolescence is getting ever more serious, says the author Peter Sandborn, who is associate professor of mechanical engineering at the University of Maryland and a member of the university's Center for Advanced Life Cycle Engineering. In the 1960s, he says, the expected market availability for chips was between 20 and 25 years; now it's between two and five. Which means that by the time a chip is deployed, it's probably already obsolete.

Could this help explain why an increasing number of telecoms network equipment manufacturers are turning to FPGAs, where the value lies not in the physical part but in the software put on it? An increasing number of systems vendors seem to be crowing about the time and money spent on developing their own ASICs — are they doing this so they can control the availability of the product? If so, that would seem to be a nail in the coffin for developers of high-end communications ICs.

Two fat wodges of paper landed on my desk earlier this week telling me that optical networking researchers Emmanuel Desurvire and Randy Giles, both formerly with Bell Labs, and David Payne from the University of Southampton, UK, have been named as finalists for the Millennium Technology Prize for their work on developing Erbium Doped-Fiber Amplifiers (EDFAs).

EDFAs, of course, are a key component of optical networks, which boost the power of optical signals without using electronics. They made the first long-distance subsea communications links possible.

Judging by the thickness of those wodges of paper, the prize organizers (and their PR teams) have put some effort into researching the nominees. All the same, a dispute seems to have arisen over on Light Reading about who really did invent the EDFA.

Important discoveries often have many contributors, and honouring just a few people is bound to put some noses out of joint. What about, for instance, the two other authors on Payne's early research paper on EDFAs, who aren't included in the nomination?

The "EDFA inventors" are competing against Sir Alec Jeffreys from the University of Leicester, UK, who invented DNA fingerprinting, Andrew Viterbi, after whom the Viterbi Algorithm is named, and Robert Langer from MIT in the US, who developed biomaterials for controlled drug release and tissue regeneration. The winner will be revealed on 11 June.

BT is behind with its rollout of its 21CN next-generation network, according to a report in icWales.co.uk. The Welsh should know, since they're supposed to be the first customers to be switched over to the new infrastructure.

The PR folk at ntl:Telewest have been quick to spot an opportunity there, and exploit it. Not only has the company commissioned a survey to show that potential customers are still "in the dark" on the timescale and benefits of 21CN, but it's also launched a website, www.showmethelight.co.uk touting ntl:Telewest's own network, which it claims is second in reach only to BT, and was built with an IP-based next-generation infrastructure from the outset.

Of course, it's all spin on ntl:teleWest's part. Not all next generation networks were created equal, and 21CN is a massively ambitious project. But I'll admit to feeling a little "in the dark" on 21CN.

File this one under "old uses for new technology". Swedish septuagenarian Sigbritt Lothberg had the world's fastest internet connection, and what did she use it for? Drying her laundry, according to Swedish paper The Local. "It was a big bit of gear and it got pretty warm," the paper reports.

It turns out that Sigbritt is not just any 75 year old, she's mother to Peter Lothberg, an optical internet guru working for Cisco. Lothberg's logic for setting up a 40 Gbit/s connection to his mum's house seemed to be "if my mother can make it work, then anyone can." There's a presentation with all the details on his company's website here.

Check out the photos, which show that's no ordinary router Sigbritt is installing in her garage, its a 4-slot Cisco CRS-1. Lothberg says his aim is to show that high-speed, long distance networking can be affordable.

With a CRS-1? Surely a tumble drier would be cheaper?

The telecoms industry's concern about the dominance of Google has reached new heights, says a report in Light Reading. Apparently, the internet content provider disclosed in a filing with the US Securities and Exchange Commission today that it has developed nuclear capabilities to help power its enormous server farms, but which could also be used to develop "land-based defence systems".

Readers, let us know of any other telecoms related hoaxes you see today, and we'll post them here for everyone's enjoyment.

Bored of life on Earth? Then you might be interested to know that Google and Virgin today announced the launch of Virgle, a jointly owned and operated venture dedicated to the establishment of a human settlement on Mars. The plan: to boldly go where no internet content provider has gone before.

Prospective Virgle pioneers are encouraged to "submit a 30-second YouTube video". If successful, they will get to join Google co-founders Larry Page, Sergey Brin and Virgin founder Richard Branson on the planet Mars "sometime in the next 20 years". The recruitment literature goes on to say:

"Sure, the work will be hard, the broadband rates low, the commodes decidedly open source, and yes, your life might be extinguished in a fiery instant of catastrophic technological malfunction. But your enriched descendants will appreciate your sacrifice, which should render worthwhile your choice to spend the rest of your (perhaps radically foreshortened) life in deprivation and uncertainty."