The SAHARA Project
By Marc Weingarten, Mon May 20 00:00:00 GMT 2002

The Wireless Global Village takes shape in UC Berkeley's labs, and may force our industry to re-think its strategy.


There was once a time in the not-so-distant past when wireless was going to fulfill the promise of McLuhan's Global Village in a very tangible, here and now fashion. The disparate strands of technology and innovation would be tightly wound in a twine of global unity that would never fray. This vast super-structure of cooperation and ingenuity would create economies of scale that would enable all of us - rich and poor, First and Third world inhabitants alike - to have access to the leading edge at an affordable cost.

But something funny happened on the way to this one-world tech utopia: People got greedy. In the wireless landscape, huge telecom behemoths sprang from whole cloth, each wanting a larger slice of the pie than the next. These companies fought tooth and nail for market share, and a handful of them got very fat indeed off of the seemingly bottomless appetite for wireless communication.

This dog-eat-dog competition is the very essence of free market enterprise, and it has netted billions for telecom companies, but according to a handful of researchers at the University of California at Berkeley's College of Engineering, who are currently working on something they call the SAHARA Project, it has also created a bloated, woefully inefficient telecom infrastructure.

Service Architecture for Heterogeneous Access, Resources, and Applications


Witness the recent financial collapse of Global Crossing, a company that banked on selling bandwidth for an abundance of fiber optic capacity, only to discover that its ill-advised spending glut created a surplus of unused infrastructure. If SAHARA had had anything to do with it, Global Crossing could have auctioned off its excess capacity, resulting in a net gain for all involved.

The objective behind SAHARA, or Service Architecture for Heterogeneous Access, Resources, and Applications, is to understand how to create efficient, cost-effective telecommunications services when those services are provisioned from multiple and independent service providers. The leaders of the SAHARA project - which includes such engineering luminaries as Randy Katz and Ion Stoica - want no less than to develop a new architecture for telecom services that supports a confederation of disparate service providers.

SAHARA is working towards improved scalability through aggregation, and the use of hierarchy and cooperation among service providers to make effective resource allocation decisions. Working with a consortium of investors that includes Ericsson, Nortel and Sprint, SAHARA wants to help mobile devices communicate more fluidly across providers. What the SAHARA researchers envision is an architecture that provides a resource management system for group policing and watching out for things like bad cell traffic overflow.

SAHARA co-founder Randy Katz is one of the world's leading pioneers in database technology and memory management. In the early 90's, he developed the concept of Redundant Arrays of Inexpensive Disks (RAID), and led the implementation of the first large-scale RAID file server for high performance applications. RAID is now a $12 billion industry, with over 150 companies marketing RAID products. In a recent National Research Council Report, RAID was cited as one of nine key information technologies that emerged from government-sponsored university research to create a multi-billion dollar industry.

In other words, when Katz talks, high-tech listens. He's long been interested in the seamless integration of various wireless applications. For his previous research project, Katz and his team figured out a way to make voice mail messages instantly translatable into text-based emails (all of Cal's Computer Science Department's research and software is freely distributed). The SAHARA Project emerged from the assumption that the wireless universe will never be homogenous.

Research Model, Business Model


"It's that whole 'Microsoft, uber alles' attitude that doesn't make sense" says Prof. Katz. "There's too much division in software frameworks. The world is not going to have a single architecture for supporting services. Everything - phones, PDAs, laptops - are going to be connected to lots of different access networks. We wanted to work towards an inter-working across lots of different access technologies inside the network."

What SAHARA strives to achieve is an architecture that allows smooth end-to-end services, even when the two parties, in Katz's words, "don't trust each other. Suppose there were Japanese tourists attending the Winter Olympics in Salt Lake City who use NTT DoCoMo as their carrier. They should be able to call up information about, say, Japanese restaurants in Salt Lake by using a Zagat-type service. That would involve a series of agreements between DoCoMo, a local carrier like Sprint, and a translating service like Babel Fish, with NTT assembling all the pieces."

The SAHARA project leaders envision a kind of brokering landscape similar to the trading markets used to buy and sell energy and other commodities. Perhaps a carrier whose spectrum capacity is straining to overflow will be able to lease, on a short-term basis, spectrum from a carrier who might be carrying too much for its current needs. Or two companies can pool their resources to build one antenna for both carriers. In other words, the current top-down, hierarchical model would be replaced by a peer-to-peer model.

"You're starting to see phone operators going after different markets," says Katz. "Virgin might be focusing on teens, AT&T may be targeting business users. They're not gonna use spectrum resources at the same time. A day trading-based environment would make sense in this context."

But as the Enron debacle has shown, trading can be manipulated if it isn't monitored properly. That is one of the loopholes SAHARA is still grappling with. "You need adequate logging capabilities to monitor the trading," says Katz. "We are very aware of that, and we're thinking carefully about making it fair and verifiable."

Democratizing the Airwaves


For Katz, the SAHARA project could lead to a greater democratization of wireless, in which the old monolithic ISP's and service carriers will have to share space in their playgrounds, or else be relegated to the margins. 802.11b Wi-Fi, with its relatively cheap hardware and low access cost, is the ideal environment for what SAHARA wants to achieve.

"We're starting to see a new hierarchy of business models that would admit more of this dynamic modeling," says Katz. "Wi-Fi employs hot spots in different places, like airports, Starbucks, hotels. Now, with companies trying to aggregate connectivity across these ISP's, you can have some kind of auction for access going on between them."

The balance of power will fall to, say, the small business owner, who can offer his access to the highest bidder. "A café owner may offer a portion of his access to different companies," says Katz. "This kind of economics-based incentive for dynamic assembling of a coverage area is more likely to happen with Wi-Fi than with something like 3G, which is more expensive and uses dedicated spectrum."

In this kinder, gentler wireless world, wireless operators would function as a kind of clearing house for assemblers: Your Sprint bill, in other words, would aggregate all of the charges for your monthly wireless services, regardless of what services you might have used. Katz says that many of the major carriers are closely watching the SAHARA project to see how they might be able to apply it to their business models. "They're trying to understand what it means to them." he says.

Looks like it could mean an awful lot, as it turns out - and sooner than they think.

It's Mobile Research Week on TheFeature! Check back daily for reports, analysis and in-depth articles about the wireless research community.

Marc Weingarten is an LA-based writer whose work appears in Business 2.0, The Los Angeles Times, Smart Business, Entertainment Weekly, The Village Voice, Vibe and San Francisco magazine.