Wireless is one of the hottest research areas among academics, who are looking at ways to make networks faster, less expensive and more energy-efficient. Here's a whirlwind tour of some of the more intriguing projects underway at schools and labs across the United States (some of which are being presented at the HotNets IV conference being held in Atlanta next week.
A better way to do municipal wireless
The meltdown of some high-profile municipal Wi-Fi projects has the industry wondering what the alternative to bringing widespread wireless access to cities might be. Researchers from the University of Cambridge in the U.K. and MIT say the secret to success might lie in exploiting the dense network of Wi-Fi access points already rampant in many cities.
In a papertitled "Architecting Citywide Ubiquitous Wi-Fi Access," the researchers introduce the idea of convincing current and future Wi-Fi hosts to grant access to legitimate guests whose traffic would be tunneled securely though the network and without hampering the host with any responsibility for it. "We offer this as an economically viable alternative to investing millions in new infrastructure," the researchers write.
"We argue that citywide ubiquitous Wi-Fi access can be architected at near-zero cost because the network infrastructure is already in place: A majority of city dwellers have a broadband connection and a personal Wi-Fi AP at home," the paper states. The researchers propose creating a cooperative of trusted Wi-Fi access points that could include implementation of gateways and servers to ensure security.
Treating wireless networks differently from wired
The debate about going with a wired or wireless network remains in full swing (See Wireless LANs vs. Wired LANs: One of Networking's 50 Greatest Arguments), but a perhaps lesser-known debate is whether to use old-fashioned wired network technologies to support wireless networks or to come up with new architectures suited specifically for wireless. MIT researchers Sachin Katti and Dina Katabi support the latter in research outlined in a paper issued in September titled "MIXIT: The Network Meets the Wireless Channel."
The researchers claim their approach can boost throughput by four times over state-of-the-art opportunistic routing schemes, which are prone to dropping whole packets if even a few bits come through incorrectly. They say these opportunistic routing schemes work fine in wired networks, but not wireless ones. "MIXIT increases network throughput by building on the inherent characteristics of the wireless medium; it embraces wireless broadcast and exploits both space and time diversities," the paper states.
Seeking more energy-efficient sensor networks
A trio of University of California at Berkeley researchers say there is beauty in procrastination -- as in networked sensors that wait as long as they can to send data back to the program or person who needs access to it. But such delays won't gratify users of all sensor networks, such as those relying on polled and schedule protocols.
In a paper titled "Procrastination Might Lead to a Longer and More Useful Life" the researchers acknowledge that loads of attention has been paid to making wireless sensor networks more energy efficient through improved operating systems, storage and communication. Their research focuses on ways to reduce synchronization costs and on exploiting the batching of data, as well as compression, without adversely affecting system users. While they find there could be some benefits to delayed communications, they also discovered challenges, such as figuring out how to establish routing schemes on the fly once sensors are waked up.
