Ubiquitous Internet Worldwide-One neighborhood at a time
Haiyun Luo
Haiyun Luo's goal is to provide free Internet access to everyone anywhere. His vision depends on the generosity of people who share this same vision and applying the peer-to-peer idea, commonly used with file sharing, to Internet access sharing. The more who contribute to effort, the better. Soon, we can all be on the Internet as easily as we can listen to the car radio. Pipedream? Not according to Luo. "It is certainly doable, but I want to make sure the system benefits everyone fairly and that it is manageable, automated, and secure."
Currently, most wireless networking involves only the first
or last hop to or from a wired network, with WiFi (IEEE 802.11) or
2/3G. "The impact of wireless is expanding," said Luo, "and it
will go beyond a hundred meters or the last mile. As we get closer
to the core of the Internet, we get lots of interesting research
problems." One of them is how to manage the somewhat limited
wireless spectrum, which spans the 3 KHz to 300 GHz frequency range
of the electromagnetic radiation spectrum. With multi-hop
wireless, problems abound, from interference, channel variations,
and security problems, to cooperative service provisioning. It is
not clear what the best way is to build large wireless networks.
The challenge is to hook people up with different interests who are
not under the same administrative unit and to get them to
collaborate, and these people have different devices on different
frequencies. Luo is looking at how to extend the popular 802.11
devices used for first and last hop to the multi-hop environment.
Several options are on the IEEE standards table: WiMax (802.16
standard) extends the hop distance from 100 feet to a mile and is
speedier than WiFi, and the blazingly fast and power-frugal
Ultra-Wide Band (UWB).
I share, you share, we all share-Everyone gains
One exciting project with huge commercial potential is called PERM, which stands for Practical End-host collaborative Residential Multihoming. PERM is a framework for 802.11 networks that allows residents to pool their Internet access with their neighbors' to improve their last-hop connectivity in a managed way and at no additional cost. It also offers resilience if the neighbors who are sharing have different types of ISPs. For instance, if someone's cable goes down, someone else's DSL can come to the rescue.
"Instead of worrying about stealing each other's signals," said Luo. "We offer a technical solution that ensures mutual benefit and security. We can verify and authenticate each other. We can encrypt the network traffic so not only will your neighbor be unable to read its content, but he will not be able to see which Web sites visit, which is not the case with the end-to-end encryption commonly used today." The current practice of access control usually disables sharing, but PERM contains ways to block particular pieces of information in a managed way so that people can share their connections with others without worrying about privacy.
Luo's product exists now and is ready to use. A 1 MB file, which can be downloaded from the PERM project page, changes the management software in the wireless router's firmware. He was able to accomplish shared Internet access with a traffic-splitting algorithm that he is presenting at IEEE Infocom in April 2006, the top IEEE conference on networking and communications. If you share it with your next-door-neighbor, you have the potential to double your throughput.
But the scope of PERM is still limited: Neighbors can only
benefit from sharing when they are around each other's homes. The
next step it to expand beyond the neighborhood. Luo did a simple
experiment by driving along Springfield Avenue, a major
thoroughfare in Champaign, Illinois, with his laptop, paying
attention to the accessible wireless points along the way. He was
amazed at how much connectivity was available. "Wireless resources
are location-based," he explained. "You install a wireless network
in your home, and that becomes your unique resource. It would be
good if there was a way to be compensated for making your system
available to the public." Luo envisions bringing autonomous
wireless network installations together into a coherent system
based on a subscription system, but not with a flat subscription
fee. Instead, he would rather that people become part of the
network in exchange for offering their own. "That way, the ISPs
will be happy, because the more people we have in our system more
customers the ISPs will have, plus it fulfills a community-wide
service, which is always good for business."
Cognitive Radio: No more wasted channels
In another research direction, Luo is working on the spectrum management problem. The wireless communication band, part of the radio frequency spectrum, is limited in range and is regulated by the FCC. A recent FCC survey revealed that only ten percent of this precious chunk of spectrum is actually in use. For example, a company like Sprint or Verizon Wireless may own a certain frequency band that is idle in a certain area because the population density is too low. Most wireless devices have fixed frequencies, which are flexible in a narrow band and are programmed in hardware, so they are unable to take advantage of unused spectrum.
Luo is working with something called Cognitive Radio, a
system that can learn from its environment so that it can
dynamically reuse available spectrum. To find available
frequencies, a user can scan the frequency band or query the system
(e.g., "I am in Champaign, Illinois. What are the frequency bands
I can use?"). A cognitive radio, on the other hand, can
automatically sense its environment by constantly monitoring the
spectrum and how it is being used and then reconfigure itself to
exploit the allocated but unused portion. To help accomplish this,
Luo is working to develop wireless protocols at various layers of
the protocol stack. "If the devices are intelligent enough," he
said, "they can pick up underutilized frequencies. Software
defined radio is just one way to get cognitive. If you have
intelligent enough hardware, you can do it in that." This creates
additional capacity and addresses the spectrum scarcity
problem.
Written by Judy Tolliver,
March 24, 2006
--
Last Modified August 07 2006 08:59:01.
