Professor David Starobinski receives Early Career Award from the US Department of Energy


Infrastructure for the really big picture. Today’s scientists manipulate mind-boggling amounts of data in their search to understand enormous questions, such as how the universe began and how living beings operate on a molecular level. And increasingly, they collaborate, across disciplines and around the world. Such collaborative research has long outgrown the capabilities of an Internet choked with spam, commerce, and virtually nonstop communicating and surfing by people from toddlers to grannies. 

To create more capacity for important science, the Department of Energy has launched a new initiative — known as UltraScience Net — that is being specifically designed to accommodate scientists working on the big questions. David Starobinski, an ENG assistant professor of electrical and computer engineering, is building a vital part of the infrastructure that will make the new scientific mega-superhighway flow smoothly.

According to Starobinski, current networks have inherent instabilities that lead to congestion: too many “packets” of information entering at the same time, parts of a network that may be restricted to information from certain sources, and programs that consume resources endlessly waiting for key decisions to be made (for example, you order a book online, your credit card information is sent to your bank for authorization, but too many other credit cards are waiting for authorization, so your purchase is held up). Currently when a network gets bogged down, various programs spring into action to try to solve the problem — to find alternate routes around the congestion.

Such ad hoc solutions won’t work in the new scientific network, Starobinski says. Instead he is working to identify the various forms of instability that occur in networks, determine their underlying causes, and estimate how likely it is that a particular instability will appear. His goal is to develop new infrastructure for the network to prevent problems from arising in the first place.

Starobinski’s work is supported by an Early Career Award from the Department of Energy. Further information about this work is available at: