A-Z Index | Directory | Careers

Special Feature: Five Questions for ESnet's Greg Bell

ESnet Helps Scientists Plug into Supercomputers, Experimental Facilities

September 6, 2013


Greg Bell

The U.S. Department of Energy operates 40 major research institutions around the country, including 25 national laboratories. The laboratories support tens of thousands of researchers at universities, laboratories and other institutions, providing them with unique scientific facilities that are powerful tools for discovery. Among these facilities are a handful of supercomputing centers housing some of the world’s most powerful supercomputers. Connecting all of these sites and facilities is DOE’s specialized research network, the Energy Sciences Network. Better known as ESnet, this network’s “backbone” can carry up to 100 gigabits of data per second, making it the nation’s fastest scientific network. Managed by Lawrence Berkeley National Lab, ESnet is led by Greg Bell.  In honor of DOE supercomputing month, we caught up with him to ask five quick questions about what makes ESnet unique.

Question: What is ESnet and how did it get started?

Greg Bell (GB): The Energy Sciences Network (ESnet) is the high-speed, nationwide network interconnecting all DOE Labs, sites and facilities. ESnet was created 27 years ago, when DOE's Office of Science combined two smaller networks into a general-purpose facility. ESnet is one of the fastest networks in the world, and in terms of the amount of traffic carried, is currently growing at twice the rate of the commercial Internet. 

Q: How is ESnet different from other networks?

GB: Commercial networks are built to carry the millions of small traffic flows generated when people surf the web or watch video online.  We sometimes call these small flows ”mice” to distinguish them from the “elephant flows” created when massive science data sets are moved around the world for analysis. Building a network for elephant flows takes careful engineering, and that’s why ESnet is different from most commercial networks. We also offer special services aimed at making our science network useful for large collaborations, and we dedicate a fair amount of time to architectural and software innovation - to keep ahead of those massive elephant flows. 

Q: What’s the connection between networking and supercomputing?

GB: Think about what happens to your laptop when an Internet connection isn't available.  It's not very useful, is it? Supercomputers are similar in some ways - because they often need to produce or consume huge data sets in the process of doing scientific work.  In fact, the DOE supercomputing centers served by ESnet have some of the fastest network pipes in the world, feeding into some of the world’s most powerful supercomputers.

Q: Many people think of networks as infrastructure or a utility, but you have a different perspective. Can you elaborate?

GB: More and more big scientific instruments are designed around the assumption that high-speed research networks will be available to connect them to supercomputers and data-analysis facilities around the world.  I like to say that modern research networks have become extensions of the discovery instruments they serve.  Networks are no longer just a collection of pipes.  If designed and used properly, modern research networks like ESnet can actually accelerate discovery for the collaborations they serve. 

Q: Your job is clearly about connectivity. When was the last time you disconnected completely?

GB: In this line of work, connectivity is an occupational hazard.  One way I disconnect is by taking open-water swims.  Thankfully, there’s no opportunity to check email when you're bobbing around in the middle of the San Francisco Bay!

About Computing Sciences at Berkeley Lab

High performance computing plays a critical role in scientific discovery. Researchers increasingly rely on advances in computer science, mathematics, computational science, data science, and large-scale computing and networking to increase our understanding of ourselves, our planet, and our universe. Berkeley Lab’s Computing Sciences Area researches, develops, and deploys new foundations, tools, and technologies to meet these needs and to advance research across a broad range of scientific disciplines.