When University of California San Diego Professor Sujit Dey looks into the future, he sees a world where “smart” wireless video clouds are filled with movies, TV shows and YouTube clips. In his vision, these clouds perform like a distributed network, invisibly feeding bits and bytes of data to mobile devices quickly and more efficiently than today’s wireless networks, effectively lightening the traffic load on wireless networks.
Professor Dey’s vision could be moving closer to reality. He and his research colleagues and students at the at UC San Diego’s Electrical and Computer Engineering Department were picked along with those of other universities around the world to explore fundamental changes to today’s wireless networks.
The multi-year, multi-million dollar research endeavor is led by Cisco, Intel and Verizon called the Video Aware Wireless Network (VAWN) project, as reported here earlier. “We now have resources to change how we see and use wireless networks,” Dey said.
Dey said that as first-generation networks evolved from 2.5 to 3G and today’s 4G technologies, advancements were primarily based on increasing capacity to meet the increase in subscribers. He sees the VAWN project as a catalyst for changing wireless networks from primarily functioning as voice and data transmission pipes.
Cisco, Intel, and industry researchers contend that demand for mobile TV and social media video is spiking the overall growth of mobile video traffic. According to Cisco’s annual Visual Networking Index Forecast, overall Internet traffic is expected to grow four-fold by 2014, with mobile video growing at over 15 times this rate.
“This research will raise consciousness and excitement about seeing wireless networks not only as links,” said Dey, “but see them as a coexistence of computing and storage.”
Dey said he believes cell towers and access points could be built up with processing power and hard drives. Rather than storing mega amounts of video content in a few CDN (content delivery network) locations, smart wireless networks would have a massively distributed disbursement of what Dey calls “smaller micro-caches,” consisting of thousands of videos or gigabytes of information.
Content would be pulled off of the Internet servers and stored intelligently inside the micro-caches associated with the base stations and access points, close to where people are accessing the video.
One thing that he and his students are exploring is how to keep fresh, in-demand video in each micro cache, like the day’s most watched videos or video-related to trending topics or even videos that are most likely to be accessed by the users in the associated cell site. This could help lower the amount of data that needs to travel across the network to get a video to play on a mobile device.
“For wireless network operators, revenue per bit of data will be low for transmitting video compared with more profitable services like texting. On the networks, video has higher and higher volume but is providing little revenue.
And Dey sees inefficiencies, like when many people in an area are viewing the same video at the same time, or even in the same hour, using their smart phones or mobile devices.
“What if those videos were stored somewhere in the networks versus being pulled from Internet servers?” asked Dey. “What if we could cache videos in the base stations or access points?”
Dey advocates building computer processing into base stations and access points, where people are tapping into their wireless service, whether it’s 3G or 4G or even a WiFi hotspot. This way, video found in the micro-caches, but in different formats, bit rates, etcetera, can be locally transcoded/transrated.
Finally, Dey envisions building a “video cloud scheduler,” which will consider the availability of multiple wireless links available to a user (cellular or Femto or WiFi), the availability of the requested video in a local micro-cache, and need and availability of processing, to schedule the resources optimally so as to increase the video capacity of the end-to-end network.
“This research project gives me a hybrid platform, almost the best of both worlds,” said Dey. “First, the university is full of young, hungry talent and their passion can get turned into creating something impactful for the industry. Second, we try but cannotalways incorporate all the practical aspects, so Intel and Cisco can influence what’s being done, how it’s being done and how the research findings can be used.”
Prior to teaching and being a founder and chief technology officer at Orvita Wireless, Dey was a researcher for several years at tech firm NEC. Dey knows what it’s like doing research inside a big company. “Not every successful research got used, but some were and there was always lots of learning of how to make research results commercially usable,” which, he said, could influence his thinking of future products and services.
Dey said the current project feels different than other industry-sponsored projects he’s led.
“The VAWN project feels different from the other industry sponsored projects because of the tight interaction we are hoping to have with Intel and Cisco,” he said.