Open Spectrum

Date December 13, 2002
Speaker Robert BERGER(Visiting Research Fellow, GLOCOM, International University of Japan)Lawrence LESSIG(Professor, Stanford Law School)MURAI Jun(Professor, Keio University )
Moderator IKEDA Nobuo(Senior Fellow, RIETI)


Robert Berger

I will talk about the idea of spectrum commons. Our vision of spectrum has been the technology that came out of the 1920s, such as radio. This technology was based on very simple receiver technology. But now we have huge amounts of signaling capabilities. So we can look at spectrum in a new way. Spectrum is inherently infinite, just as there are infinite points along a line. It is the technology that allows us to use the spectrum.

We start to see the scope of what we can do with spectrum with such technologies as the 802.11 wireless LANs, where the technology uses less spectrum, and therefore you can have hundreds of people using wireless LANs, in an office for example, and they are getting 10 megabit per second throughput. And by shrinking the size of the coverage of any radio, you can utilize spectrum much more.

Just recently, the US FCC issued a spectrum policy task force report. It is amazing because the report is raising issues that the agency would have not raised just two years ago. One suggestion they propose is the low power underlay. By using techniques such as spread spectrum, ultra wide band, and other technologies, we can take what used to be considered noise and use it for actual signals. The second technique is to reuse idle spectrum.

One of the concepts the FCC came up with is interference temperature. This is a way to measure how much interference is being generated in an area. So you could have regulations that would say a device would be allowed to operate, using, say ultra wide band, in an area where you are below a certain interference temperature. This allows us to create an underlay commons. Spectrum can become available for new devices that will not interfere with the existing use, so you can underlay uses and potentially have millions of low power devices operating at the same frequencies as existing legacy devices. This is what we call a commons.

The second technique has several names: the agile radio or software defined radio. This is a radio that can take advantage of the signal processing that we can now implement in semi conductors. It can sense and when a prime user starts transmitting in that frequency, it would then release and use another chunk of frequency. It can scale its power based on its application, the local conditions, and the rules set by local authorities. This opens up huge amounts of capacity, and, effectively there is infinite spectrum if we use it correctly. This is the commons.

Lawrence Lessig

The conversation we are having today is equivalent to a conversation we might have had in 1982, when technologists would have come before you and said, "This new idea of packet switched networks could enable something we might call the Internet." Most people in 1982 would have said, "I don't get it. Why is this important?" The lesson is that we non-technologists must listen when the technologies say, "There is something extraordinarily important that is just about to happen. It will have an extraordinary impact on social resources like spectrum." And that is what is happening in the context of spectrum.

We need to understand the costs of a property system. We need to understand why they will be costly in the context of spectrum. We have a way to think about spectrum that is intuitive and wrong. We imagine it as a thing out there like coal and we ask, "What is the best system for allocating this thing?" We ask this as if there were a fixed quantity of this thing out there. That intuition is wrong. According to the old spectrum technologies, this was the way for using the spectrum to communicate. The old receivers got confused very easily if there were two transmitters close by. Televisions have been too stupid to hear the difference between, say, channel 5 and channel 6.

There are two gains for utilizing spectrum when you use broadband. The first is processing gain. Given a certain amount of communication possible for a transmitter, the gain would be tricks that we can build into the system to make it so that more information is communicated. With computer technology, we get a processing gain, which is the intelligence needed to distinguish which bits of data a device should be gathering and putting together. The broader the bandwidth, the greater the processing gain becomes.

The second type of gain is cooperative gain. Instead of there being certain access points, each user in a given space would act as a node in the network, so that your communication does not have to jump all the way to the access point. The hopping between peers would lead to the main access point. If we can reduce the distance between any two hops, the power needed to hop is much less. Therefore, the aggregate capacity increases dramatically. And the gain is produced by the cooperation of these nodes. This means that when you increase the number of users, you increase the capacity of the system.

This is creating weirdness in economics because there are very few situations where increasing the number of users increases the capacity. It is a network effect, but it has a profound effect on how we should think about allocating resources. The property system is designed to allocate the resource to the highest value user. The cost of the property system would be that it might reduce the capacity of the system. So the policymaker has to figure out which is more important: increasing the capacity of the system or allocating the resource to the highest user.

Property can reduce the capacity of the system because it effectively narrows the bandwidth that is available for this spectrum system to use. If you sell off chunks of the bandwidth for license, there is less bandwidth available. We are just seeing the beginning of the spectrum commons. The real gain will come when we move to a mesh-architecture when each receiver is a node in the system. Property systems only make sense in places where there are capacity problems. For FCC Chairman Powell to be successful in liberating spectrum from inefficient use, we need policy to protect common space.

Jun Murai

When Internet technology advances further, spectrum will become a critical issue. We have to find a way to share this space effectively. To this end, you may want to synchronize the time element of spectrum or to detect the empty spaces to make full use of spectrum. To avoid interference, you must find ways to avoid conflict. This presents us with technical challenges.

Questions and Answers

Q: For those who support open spectrum, what are the incentives that you are willing to give to the incumbents to open spectrum further?

Lawrence Lessig: It is unknowable whether the powerful forces that are in place now will step back enough to allow commons to take over. But the strategy should be to go slow with the deployment of the spectrum commons technologies, let the market grow as fast as it is growing and allow the FCC to push it along, and to use strategies that everybody intuitively understands, such as the white space strategy (where one user would use space that is idle). It would be hard for the incumbents to defend any resistance to the white space strategy. The strategy is to chip away at the exclusive use model and to get people to realize that the fastest growing model out there is an industry that depends on spectrum commons.

Robert Berger: Having a commons does not mean that anything is possible and everyone can do whatever they want. I still think there will be a need to regulate the equipment. The equipment manufacturers must certify that the equipment does not exceed a certain power and that there are certain rules that limit how much power you can emit.

Jun Murai: It is true that a feeling of concern is healthy when there is conflict. When you use wireless LAN, for example, you cannot interfere with a microwave oven. If you turn on the oven, the wireless LAN corrupts and you will not be able to connect.

Lawrence Lessig: Yes, there will be rules, but the rules will apply to the devices, not the use of the spectrum. And we will have a future based on property, but I hope it is property based on devices and not on spectrum. But there is something to worry about when it comes to rules that govern devices. The best way for an incumbent to screw up this evolution is to get "etiquette rules" enacted that benefit their particular vision and could basically destroy the competitor.

Robert Berger: You hear the FCC worrying that there will be chaos and that the people in the commons will rise up and try to take over. I am more concerned about the opposite: if you give property rights to incumbents, they will go into the courts and stomp on the commons use.

Q: IIf we go back to 1959 and assume spectrum is a scarce resource. The Ronald Coase wrote a paper that criticized the hierarchical coordination system to allocate spectrum to support the property based market mechanism. Would you have supported that conclusion?

Lawrence Lessig: In 1959, Ronald Coase was right. His conclusion about property in 1959 was correct given the technology of the time. It is not because spectrum is a fixed resource; it is that the technology at that time could do very little about it because to get the cooperation gain you need the computing power, which in 1959 you did not have. Over time, Coase's conclusions were a function of technology.

Q: In this new ubiquitous world, how do you allocate without creating conflict? Perhaps the architecture itself will create larger problems. Perhaps the green-field method (unregulated) is a way of opening up the spectrum.

Lawrence Lessig: If you are talking about IP addresses, I do not see this as a problem, if we move to Ipv6 because there would be no coordination problem with the address. The nodes in a network do not necessarily need to coordinate with a central authority to be part of that network. With the 802.11 technology right now, it creates an arbitrary IP address for all devices that are communicating that is only recognized in the local network. The local network is doing the coordination and then plugging into the public IP network. So we could set up a local wireless network that used all sorts of IP addresses. As long as it was plugging into a translator that was then connecting that to a public network according to the coordinated IP addresses, there would be no problem with the coordination of those addresses. And Ipv6 would assure an ample supply of addresses.

Q: Internationally, what will be users demands in the evolution of open spectrum?

Robert Berger: Users exercise their interest by purchasing devices. We have to move regulation along so that people will invest in the technology, which will create demand.

*This summary was compiled by RIETI Editorial staff.