Scaling Bandwidth to Fit the Job -- in Real Time

MIT's John Wroclawski explains how his personal router would let wireless users get the level of service they need, when they need it

Fed up with your cell-phone service? Heck, you would have bailed long ago, except you're locked into a long-term contract, right? Disgruntled dialers, take heart. Massachusetts Institute of Technology research scientist John Wroclawski is working on a technology and business model that could free mobile-phone users from being locked into a single network.

Wroclawski has invented a little box he calls the "personal router." It would allow customers to hook their cell phones into small-scale wireless local-area networks to make phone calls that access the Internet. So customers could roam from a slower network to a faster one, depending on what level of service they required. If you need to see only text, the router would chose a slower-speed service. Pictures? You would switch to a higher speed.

Whether cell-phone companies would agree to participate in a such a scheme remains an open question. But they may not have a choice, as wireless LAN coverage improves over time and begins to rival cell-phone coverage. Wroclawski recently spoke with BusinessWeek Online Correspondent Olga Kharif about his project and other aspects of emerging wireless technology. Here are edited excerpts from their conversation:

Q: How would your model and technology work?


The thing that we are trying to make possible is called the personal router. The personal router is both an idea and a little box. The little box, wherever you are, looks around, sees what services are available, and, based on an understanding of availability and what you are trying to do at that instant in time, negotiates on your behalf for service.

Q: How would the personal router make the choice?


This little box would chose a service based on an understanding of what you were doing. So if you were just reading e-mail, it might choose the slow service. But if you were browsing the Web, it might choose the faster service.

Or in fact, it might be more sophisticated than that. If you were just reading text articles on the Web, it might choose the slow service. But if you decided to download a picture, then it might, just for a brief moment, move to the faster service, because you might be willing to pay a little bit more just to download that picture fast. So it's a very dynamic negotiation, based on an understanding both of what services are available and what you are trying to do.

Q: How would the billing be done?


If the box has chosen the service for which you have some kind of a long-term agreement in place, then the billing is done the normal way, through a monthly bill. If it has chosen a service that is being provided as a loss-leader or as an adjunct to some other service that you bought, like buying a cup of coffee, then there is no bill. If it has chosen a service that actually has a per-minute charge or a not-flat-rate charge of some kind, then there are a wide variety of choices.

If you had some kind of a credit-card scheme that could actually handle very small payments, it might bill your credit card. But that's probably impractical, because credit-card companies tend to want a $10 minimum per transaction. But there have been a lot of people who have looked at ways to efficiently transfer very small amounts of money from one person to another -- electronic cash and things of that sort. So that's a possibility. There's another whole class of possibility that we think is extremely interesting, which is to interject a third-party broker into this path.

Q: How would this third-party broker work?


You might pay $19.95 a month, plus some usage charge, to a third party, who would then aggregate all of your payments with other payments, take some of the risk, and then pay out money to people who actually provided services. It's kind of an HMO analogy, where instead of paying each doctor separately, you pay a certain amount of money for health care. The HMOs kind of negotiate on your behalf to get the best service -- or best, according to some definition.

Q: Who would play this HMO role?


One group of people that is very well situated to play that role is the wide-area service providers [such as Sprint or Verizon]. In one sense, this project is actually kind of threatening to them -- you know, you roll out a whole alternative business model for wireless when they've just spent $4 billion on 3G [third-generation] wireless spectrum.

But on the other hand, it's also a very big opportunity. They can, perhaps, handle the billing and business relationship with you, and be kind of a trusted intermediary. And [they could] use not only their own networks but get short-term access to other people's networks.

Q: Wouldn't the wireless operators somehow lose out on the money they are getting if a wider range of people would be able to get into the field?


There are three possible answers. One is "yes." The second is that even if the market stayed at the same size, but the operators continued to play a role in it -- just a different role -- then, in fact, they might not lose out. Instead, what you might see is that they would continue to get a fair bit of revenue and their capital costs would go down. Or a third possibility is that the market actually grows to the point where they may lose out in proportionate terms, but that since the absolute market is bigger, they do fine anyway.

Q: Wouldn't increased competition from wireless LANs be a threat to the carriers who bet billions on 3G?


There's great value to a seamless service that doesn't work just in your living room, but that actually works wherever you are [the way 3G would]. The kind of mixed technologies [combining various networks] that I've been talking about are another way to accomplish it. They may well be noticeably cheaper, overall. They may have performance advantages.

I think that the fact that there is a sort of technical, different approach evolving, is another reason that perhaps carriers that bet on 3G overestimated the market -- or at least how near it is to returning any real revenue.

Q: Given the wireless LANs' existence, why would we bother doing 3G or 4G networks?


I think it might play out that way. It seems possible. It may be that if wireless LAN technology becomes much more widespread, you may see that it will drastically slow the deployment of 3G, if nothing else.

Edited by Douglas Harbrecht

Before it's here, it's on the Bloomberg Terminal.