When you first ask computer science professor Anthony Joseph about his research into mobile telephone services, he instantly launches into an imaginary story about a Japanese tourist, Ms. Tanaka, visiting Salt Lake City to attend the Olympics.
Back in Tokyo, her cellular service is provided by NTTDoCoMo. But thanks to a roaming agreement with Sprint, she can make and take calls even when she's in Utah. In Japan, Tanaka frequently uses NTTDoCoMo's restaurant recommendation service for her mobile phone to pick a spot for dinner. In the U.S., Sprint offers a similar service in partnership with Zagat's Guide. But the reviews are in English and not formatted for display on Tanaka's Ericsson handset. Fortunately, JAL Travel put together a special service just for Japanese tourists attending the Olympics. With the help of several other companies, JAL takes the information from Zagat's, translates it into Japanese, and reformats it for display on Japanese handheld information appliances. In moments, Tanaka signs up for a two-week service agreement with JAL TRAVEL and the charge appears on her monthly NTTDoCoMo bill.
Making that fantasy a reality is a massive technological and logistical undertaking. And that's precisely what makes it an ideal research project, says Joseph, who is collaborating with professors Randy Katz, Ion Stoica, and a dozen graduate students.
"We each bring a slightly different point of view to the problem," Joseph explains. "That's something very unique about Berkeley. It is the rule, not the exception, that we work together on these kinds of projects."
Launched last summer, the Service Architecture for Heterogeneous Access, Resources, and Applications (SAHARA) project is poised to reinvent nearly every aspect of mobile telephone service. The aim, Joseph explains, is to topple the monolithic structure deployed by a single business entity and replace the "all things to all people" model with a dynamic confederation of service providers cooperating as needed.
The first thing the researchers determined about today's wireless telecommunication networks is that they're amazingly inefficient and unnecessarily expensive. For example, one popular service provider's wireless spectrum resources may be overburdened, resulting in bad service, while another may have plenty of spectrum to spare. Why couldn't one provider sell capacity to another as needed? Additionally, cell antennas are often unnecessarily duplicated. In the SAHARA scheme, an antenna "operator" could provide the infrastructure for all the service providers. Finally, perhaps all service providers could share a better-provisioned common link to the Internet.
Identifying the problems is the easy part though. A "service composition"
architecture requires complex monitoring and measurement of available
resources along with new auction-like models for buying and selling
those resources, Joseph explains. For example, during a high demand
period where wireless service is at a premium, at the Olympics
for example, a provider may be willing to drop big bucks to satisfy
its customers. (Of course, that price increase would then be passed
on to the end user.) Still, even if the resources are available,
the mobile services need to be accessible from a wide range of
devices.
"Clearly, an application that runs fine on my high-bandwidth well-connected laptop won't run as well on my mobile phone," Joseph says. "We're working on ways to adapt applications on the fly."
In the end though, the successful deployment of a SAHARA-like system depends
on trust. And trust management in a massively competitive marketplace
like wireless telephone service is tricky business. Currently,
relationships between service providers are based on carefully
detailed contracts between big-name commercial enterprises. But
in the SAHARA model, the number of providers and relationships
massively increases.
"There needs to be a way to make sure that when you make a phone call the right person is billed for the right amount and that you get the quality of service you expect," Joseph says.
One solution, the SAHARA researchers suggest, is to enlist a trusted auditing service, similar to a better business bureau for wireless, that would filter out unreliable providers.
Joseph has no shortage of SAHARA spin-off projects in mind.
Imagine an optional location-based service that could notify you
if someone on your "buddy list" is in a café around the
corner, he says. But while the researchers are eager to present
a big picture vision for SAHARA, they're also aware that progress
takes time. And even incremental innovations, Joseph says, could
make the wireless world more inviting and accessible to end users.
"We want to propose the grand vision, but we also want to say how we get there from here," he says. "So we're considering how technology that could be deployed in five years might be integrated with legacy technology. Even if we come up with the world's greatest system, it still has to work with that old rotary phone."
SAHARA Project
