November 25, 2009 –
The explosion of new Internet applications is fueling radical change in the networks that support them. Last week’s Igniting Technology event attracted nearly 150 guests interested in learning just how that evolution is progressing and what business opportunities it is spawning.
The 10th in the Igniting Technology series, “Network Generation: Transforming the Global Infrastructure to Address Evolving Demands,” featured three UC Irvine researchers, the CTO of chip manufacturer Broadcom and a venture capitalist, each contributing his unique perspective.
The evening was moderated by Michael Guiliana, partner at Knobbe, Martens, Olson and Bear, the intellectual property law firm that supports the series. Guiliana advised future entrepreneurs in the audience to build a bifurcated business plan for their startup companies, one that focuses on strategic acquisition possibilities as well as new products and services. “Develop things that are easily adoptable by strategic partners. It’s a way to bring in cash and ensure early success and momentum for your entity,” he urged, citing CALIT2 spinoffs HIPerWall Inc. and CODA Genomics as good examples of that strategy.
Researcher Paul Dourish, from UCI’s Laboratory for Ubiquitous Computing and Interaction, shared his views on computer ubiquity. He explained that while ubiquity can make our lives easier, sometimes it presents complications and challenges. “I always say ‘we have find ways to get things to talk to each other and then find ways to make them stop,’” he quipped.
Ubiquity also creates interesting juxtapositions. He told the story of an aboriginal woman teaching native practice, culture and language through a university program that employed Internet technology to connect her with classrooms in metropolitan Australia.
Dourish and several other program advisors were on a Skype call with her when one suggested she teach a particular lesson by having a conversation about the topic with her brother. She explained why that was impossible: in her culture it is taboo for brothers and sisters to converse with each other unless they’re mediated. Also, she said, the location from which it was suggested she teach the lesson was hallowed ground, preventing her from sharing the story from there.
She could teach the lesson from elsewhere, she said, but there wasn’t good 3G coverage available, so they would have to use a SAT phone, which would cause delays that would interfere with the Skype call.
“It was a beautiful little moment, the intersection of two infrastructures of communication,” said Dourish. “One of them is a technological infrastructure that involves 3G and satellite networking, and at the same time, a cultural infrastructure that says who is allowed to tell what stories where, to whom, in what ways and under what circumstances.”
Ubiquity creates the need to distinguish one place from another, Dourish continued. As the number of mobile technologies and interconnected devices grows, “infrastructures of distinction” become more important. “It’s important to create not only interconnectivity,” he said, but also ways to differentiate “my house from your house, work and home, public and private space. What we want technology to do is create for us patterns of difference, ways of distinguishing one zone from another.”
This applies globally as well. Referring again to the aboriginal teacher, he said, “One of the challenges in creating worlds of global connectivity and ubiquitous computing is to understand how to bring things together and make them the same, while at the same time making visible the boundaries in a world that has become available to us because of technology.”
Security and privacy are also significant issues, according to Gene Tsudik, managing director of SCONCE, UCI’s Secure Computing and Networking Center. The Internet has caused sensitive information to become increasingly available to the public. “There’s more and more transfer of sensitive information but there’s less and less accountability, and clearly, less and less privacy,” Tsudik said.
Along with his colleagues, he is developing techniques that allow two parties to share information without jeopardizing privacy. He discussed several examples with the audience, including one in which the Department of Homeland Security seeks to compare an airline’s passenger manifest against its terrorist watch list.
There are currently two choices, he said. The airline and or DHS could turn over their complete lists to each other, or both could refuse to share information.
He is developing a private set intersection (PSI) ‘black box’ approach that will allow DHS and the airline to compare information in such a way that only the intersection of both data sets is identified. If one name for example, shows up on both lists, that name will be disclosed, while all of the other names on both lists remain private.
Progress is slower when it comes to spam, phishing, viruses, worms and other annoying and dangerous Internet schemes. “I wish I could tell you some good news but there is no silver bullet. It’s only baby steps, and it’s an eternal battle of us vs. them,” Tsudik said.
The next speaker, Broadcom Enterprise Network Group CTO Nick Ilyadis, discussed how the Ethernet is enabling worldwide broadband convergence, which he described as the ability to deliver any content over any media.
The emergence of the miniature superchip, which holds computer, video-display device, touch-screen interface and controller, is creating an even greater demand for more bandwidth. “We’re all connected, all the time,” Ilyadis said, adding that current 3G networks and soon-to-be-released 4G networks are meeting some of the demand as broadband usage continues to increase. A 4G network carries 10,000 times the bandwidth of a 2G network, he added.
Ethernet allows data packets to travel end-to-end from computer to computer without being changed into cells or bits. Carrier Ethernet is now available in large apartment buildings, corporate headquarters and for consumers.
It’s a unifying technology, he explained, for economic and performance reasons. It gives service providers an elastic network that is more flexible when usage increases. It offers the broadest range of speeds, from 10 MB/sec to 100 GB/sec. And it has taken on the required functions of service providers.
But it was not designed to run on a carrier network, so chips have to be devised that can manage traffic flows and time synchronization. Ilyadis is confident that those chips are the future of Internet communications. “In the last 10 years we’ve gone from 500 MB on a piece of silicon to 500 GB. The Ethernet has integrated a tremendous amount of bandwidth onto a single device.”
Convergence is also a theme in quality-of-service discussions. Computer science professor Scott Jordan said the four types of networks – Internet, cellular, cable and regional long-distance carriers – never realized the extent to which they would all be used for purposes different from those for which they were originally built. “Now, they’re all starting to look about the same,” he said, carrying traffic from email to voice to video streaming.
Quality of service has begun to suffer as applications proliferate, necessitating a new approach. Jordan said there are two possible models: a reservation paradigm that reserves bandwidth for limited, selected streams, a system similar to a freeway toll lane; and a prioritization paradigm that gives priority to selected material or providers, much like the postal service’s priority mail.
Eventually, this should produce an Internet that’s capable of delivering everything consumers want, in a timely manner. But who’s going to decide which approach is adopted and which streams or service providers will get priority? “We’re going to see a lot of competition but it’s unclear how this is going to play out,” he said.
Congress and the Federal Communications Commission are looking into implementing Internet regulations that could provide answers.
While there are laws regulating telephone networks, wireless networks and the like, there are no laws at the federal level about whether or how to regulate the Internet “You’ll see this in the headlines over the next several months,” Jordan said. “There’s a huge debate in Congress and in the FCC right now.”
Craig Gunther, managing director of Blade Ventures, brought a venture capitalist’s perspective to the conversation. He told the audience that by 2013, annual IP traffic will be measured in zetabytes instead of gigabytes. (A zetabyte is 1021 bytes.) IP video will drive the increased traffic, while wireless communications will comprise a smaller piece of the pie, although it will remain a crucial component.
“You have to understand your market,” Gunther told the hopeful entrepreneurs in the audience. “Your solution doesn’t have to address all [market issues] but you need to understand where it will fit in.”
Service providers are under intense competitive pressure, he said, echoing previous speakers’ assessments. “In serious demand issues, innovation is the answer.”
He added that social media and mobile applications – most of which aren’t paying their own way – are in their early stages but they have the potential to accelerate new business models. “Some of the big money makers down the road will be the ones who figure out [how to monetize social network sites],” he predicted.
Other promising areas of interest for venture capital include: more efficient utilization of existing infrastructures; management and monetization of infrastructure by providing value-added services; smart sensor networks and smart grid technologies; and wireless healthcare devices and applications.
The presenters then answered audience questions, after which participants enjoyed a buffet dinner and viewed several CALIT2 demonstrations.
View presenters’ presentations:
Paul Dourish, “Space and Seams in Ubiquitous Computing”
Michael Guiliana, “Network Generation Overview”
Craig Gunther, “What’s Driving the Change?”
Nicholas Ilyadis, “Broadband Convergence and the Ethernet as the Unifying Technology”
Scott Jordan, “Challenge in Internet Multimedia: Technology, Economics and Law”
Gene Tsudik, “A Snapshot of Some Security Research at UCI”