Broadband on Rails

http://www.technologyreview.com/
By Rachel Kremen on Tuesday, December 30, 2008

A compact lens could make high-speed Internet access commonplace on trains.

Internet access can make a train trip far more productive and enjoyable. But train-mounted satellite dishes that send and receive data can't be used on a lot of routes, as the standard hardware is too big to fit in some tunnels. Now researchers at the University of York, in England, have developed an alternative: a dome-shaped plastic lens that's less than half as high as a typical satellite dish. The system, which was developed with funding from the European Space Agency, is also designed to track multiple satellites at once, making it more reliable than a dish.

"Here in the U.K., a lot of our railway infrastructure is very old," says John Thornton, a research fellow in the Department of Electronics at the University of York, who led the lens research. Low bridges and tunnels offer minimal headroom for satellite dishes, which Thornton says are about 62 centimeters high. Thornton's lens, in contrast, is only 30 centimeters high--short enough to meet the needs of the train industry.

The York project is based on an existing design, called a Luneburg lens. "The traditional approach would be to make [the lens] out of novel materials with certain properties," says Thornton. "I thought, 'What materials are practical and could work?'" Ultimately, the team decided on the plastics polyethylene and polystyrene, which are less expensive than the materials traditionally used to make Luneburg lenses but achieve the necessary performance. Thornton says that recent laboratory tests confirmed that the lens was able to receive digital video broadcasts, meaning that it could handle at least four megabits of data per second.

The York system also offers increased reliability. With a traditional satellite system, a separate dish is required for each satellite, and the whole dish has to move to track the signal. Moving an entire dish is fine if it's mounted on a stable structure, such as the roof of a house, but not if it's affixed to the side of a train that's running through tunnels and under bridges. A lot of room is required around the device at all times, to ensure that it doesn't hit something while tracking a signal.

With Thornton's device, incoming radiation bounces off the surface on which the lens is mounted. The lens concentrates the reflected radiation to a single point on its surface, where it's collected by a motorized antenna called a feed. To track the signal, only the feed needs to move, as opposed to the entire dish in a conventional system. Moreover, several feeds can roam around the surface of the lens at once, collecting signals from satellites in different locations.

Having extra feeds increases the redundancy of the system, Thornton says. "If one of the possible feeds isn't working, then you've got a spare." Different beams could also be enlisted for different services, he says, noting that one could be used to provide live television while another is used for Internet access.

Ratul Mahajan, a researcher with Microsoft's networking group who has been working on wireless Internet connections for cars, questions why Thornton chose to use satellite Internet instead of 3G, a telecommunications standard that's becoming common in cellular-telephone networks. "Why use satellite at all?" Mahajan asks.

Thornton says that 3G currently doesn't have the kind of geographic coverage required for continuous Internet access along train routes. Upgrades to the cell network, he says, tend to be concentrated in towns. "Each base station can only offer the highest data rates to users typically one or two kilometers away, so a truly vast number would be needed to cover all the railway routes in a country the size of the USA, or even France," Thornton says.

Thornton is currently trying to find a commercial partner for his system but admits that it's not ready to hit the rails just yet. In fact, it has yet to be tested on a moving vehicle. The team still needs to develop a control system and protocols for handling multiple satellite feeds.

Software development predictions for 2009

http://www.infoworld.com/
By Neil McAllister on January 2, 2009

A battered economy will mean tightening belts, changing customer allegiances, and the Web as the platform of choice.

What would the New Year be without that time-honored publishing tradition: predictions? By the time you read this post it will be January already, but as I write it the champagne corks have yet to pop, so there's still time for me to gaze into my crystal ball. Right or wrong? Only time will tell. But here are a few thoughts as to what 2009 might have in store for the software development market.

Microsoft struggles to retool its image
Expect more overtures from Redmond toward the open source community as Microsoft continues to try to soften its image as an anticompetitive thug. We may even see an interesting Microsoft product or two running on Linux by year's end.

Interest in Windows Azure and related cloud computing technologies will continue to grow throughout 2009, but Microsoft should watch its step. The meaning of "cloud computing" is opaque enough already. Microsoft shouldn't muddy the waters further by being too liberal with ambiguous marketing terms such as "mesh" and "live," the way it did with .Net.

Also, Microsoft's "software plus services" concept is a strong direction, but it needs to be sure that its developer ecosystem comes along for the ride. As Steve Ballmer so famously observed, Microsoft achieved much of its success through the hard work of its ISV partners. Those partners need to be reassured that they won't be reduced to sharecroppers in Microsoft's service-oriented vision of the software market.

But of course, the big distraction next year will be Windows 7. If Microsoft can't get its OS house in order, it's going to have a hard time convincing developers of the value of its new technologies. More than Ray Ozzie's vision of the future, the company's ability to hold its core products together will be the ultimate test of Microsoft Anno Bill.

Java moves toward an open source mindset
Oracle's acquisition of BEA Systems made it one of the largest players in the Java application server market. But it's still too early to say how BEA customers have weathered the transition. Some might not appreciate their contracts being subsumed into Larry Ellison's software juggernaut.

On the other hand, the Red Hat/JBoss merger has proven to be a comfortable match for most JBoss customers, the majority of whom were Linux users to begin with. And Red Hat shows a strong interest in Java; for example, it has put considerable effort into the IcedTea project, a fork of OpenJDK that improves upon Sun's open source Java stack.

The major proprietary app server vendors are likely to continue to dominate the market in 2009, but expect the movement in favor of 100-percent open source Java stacks to grow as the year rolls on. As awareness of Java's new, open source status spreads, expect developers to view the technology in a different light and begin choosing their vendors accordingly.

Twilight for Sun Microsystems?
Pity poor Sun. Floating the idea of its software installed base as a marketing channel was a gutsy play, but in light of current economic conditions it was woefully mistimed -- as so many of Sun's recent moves have been.

JavaFX is interesting technology, but it's too arcane and far too late to the party to become a serious competitor to Adobe or Microsoft. And as a mobile platform it's stillborn; Adobe has struggled for years to make Flash a major player in the U.S. mobile space, to little success. By comparison, a Sun-dominated smartphone applications market is a pipe dream -- particularly given Apple's well-documented disinterest in Java for the iPhone.

No, my fear is that 2009 may see the beginnings of the Big Cataclysm for Sun. Executive change could be in the cards, but what top talent would risk taking on such a toxic company now? Sun's best option may be simply to sell itself off -- piece by piece, if it has to. Big Blue already does as good a job of marketing and selling Sun's technologies as Sun does. Maybe Sun could become a subsidiary of IBM Labs?

The Web loses its version number
When budgets are tight, the best option is to make use of the technologies you already have and the ones you can get for free. For today's business applications, that means the Web, which will continue its rise as the dominant development platform.

While the Web itself won't lose any popularity, however, expect a little less emphasis on the "Web 2.0" moniker in 2009. For starters, it hardly means anything anymore -- if it ever did. But AJAX and related technologies still haven't proven their value for a lot of enterprise applications, beyond adding UI flash. What Web 2.0 capabilities do make their way into business applications will be the result of open source toolkits such as Dojo, while proprietary products are likely to fall by the wayside -- and that means the marketing hype will cool down a little, too.

And I expect the same will be true of Microsoft's Silverlight. Microsoft isn't likely to woo many significant partners to the platform when Flash's reach is so much wider. Adobe wisely began trying to bridge the gap between designers and developers years ago, and its efforts will keep paying off in the New Year. While developers will continue to pay Silverlight lip service, 2009 will show it to be the HD-DVD of the Web: a nice idea, but not the one that wins.

The economy looms large
Ultimately, economic conditions will play a huge role in how 2009 pans out for the developer community. When customers aren't buying, tool vendors don't innovate -- so don't expect many groundbreaking new technologies to debut this year.

Among enterprise customers, tightening budgets are likely to put the kibosh on many an ambitious new project. But smart companies will realize that process automation is one of the best ways to reduce costs in any business. Now may actually be the ideal time to revisit old software schemes that got shelved back when staffing budgets were flush. Layoffs and hiring freezes will mean there are fewer developers to go around, however, which could make smart projects infeasible for the time being.

On the plus side, while offshoring of development projects will continue among larger companies with well-established overseas divisions, I predict emphasis on offshoring will decrease among smaller companies and first-time customers. Smaller shops won't want to look like villains during the downtimes, especially when local talent is abundant in and willing to work cheap.