Project Connecting Hyades Supercomputer Cluster with NERSC Recognized with CENIC Innovations Award
The project connecting the University of California Santa Cruz’s Hyades Supercomputer Cluster to the Lawrence Berkeley National Laboratory’s National Energy Research Scientific Computing Center is being awarded the CENIC 2018 Innovations in Networking Award for Research Applications.
Project leaders being recognized are members of the University of California Santa Cruz (UCSC) Astronomy and Astrophysics Department: Piero Madau, Joel Primack, J. Xavier Prochaska, Enrico Ramirez-Ruiz, and Stan Woosley; members of the UCSC Science DMZ team from UCSC Information Technology Services: Shawfeng Dong, George Peek, Joshua Sonstroem, Brad Smith, and Jim Warner; Peter Nugent at the Computational Research Division at Lawrence Berkeley National Laboratory; and John Graham with the Qualcomm Institute at the California Institute for Telecommunications and Information Technology (Calit2).
Astronomy and astrophysics are disciplines that require processing massive amounts of data. A single night's survey of the sky with a state-of-the-art telescope can yield a tremendous amount of data, which is often analyzed in real time. Today, nearly all scientific research and data analysis involves remote collaboration. To work effectively and efficiently on multi-institutional projects, researchers depend heavily on high-speed access to large data sets and computing resources.
UCSC responded to this challenge by organizing a unique collaboration between scientists and technologists. Several units within UCSC’s Information Technology Services worked with UCSC researchers to win an NSF grant to establish a campus Science DMZ. A Science DMZ is an architecture developed by the US Department of Energy’s Energy Sciences Network (ESnet) to support faculty and research projects. The UCSC Science DMZ offers a 100 Gbps network connection between UCSC and participating institutions.
This effort dovetails with the Pacific Research Platform (PRP), now in development by researchers at UC San Diego and UC Berkeley in collaboration with CENIC. The PRP integrates Science DMZs on multiple campuses into a high-capacity regional “freeway system” that makes it possible to move for large amounts of data between scientists’ labs and their collaborators’ sites, supercomputer centers, or data repositories without performance degradation.
Using PRP, UCSC connected its Hyades Supercomputer Cluster to the Lawrence Berkeley National Laboratory’s National Energy Research Scientific Computing Center (NERSC) over CalREN, CENIC's 100 Gbps optical network, and over 100 Gbps peering with ESnet providing connectivity to NERSC. Peter Nugent, an astronomer and cosmologist from the Computational Research Division of LBNL, was pivotal in this effort. This connection enables UCSC to carry out the high-speed transfer of large data sets produced at NERSC, which supports the Dark Energy Spectroscopic Instrument (DESI) and AGORA galaxy simulations, at speeds up to five times previous rates. These speeds have the potential to be increased 20 times the previous rates in 2018.
“To accelerate the rate of scientific discovery, researchers must get the data they need, where they need it, and when they need it,” said UC San Diego computer science and engineering professor Larry Smarr, Principal Investigator of the PRP and director of Calit2. “This requires a high-performance data freeway system in which we use optical lightpaths to connect data generators and users of that data.”
Initially, when Hyades was connected at 10 Gbps to the campus production network, data transfer was slow and cumbersome. Then, in March 2017, PRP provided a FIONA box to facilitate data transfer between UCSC and NERSC. A FIONA box, or Flash I/O network appliance, is constructed out of commodity parts. FIONA boxes are highly optimized for data-centric applications and act as "data super-capacitors," increasing the possible bandwidth speed to 40 Gbps or greater.
Computational astrophysicists at UCSC now regularly use the supercomputing resources at NERSC and routinely transfer terabytes of data between Hyades and NERSC. For example, the enhanced speed has greatly facilitated work by cosmology researchers Piero Madau and Joel Primack, who have been using supercomputers to simulate and visualize the evolution of the universe and the formation of galaxies while comparing the predictions of these theories to the latest observational data. The new connection also supports astrophysicists Enrico Ramirez-Ruiz and Stan Woosley, who use supercomputers to simulate violent explosive events like supernovae and gamma-ray bursts.
Brad Smith, Interim Vice Chancellor of IT Services at UCSC and PI on the NSF grants used to fund the Science DMZ work, said: “The ITS division at UCSC is thrilled to be able to facilitate this important astrophysics research. With funding from NSF we were able to build a 100 Gbps connected Science DMZ, and through collaboration with CENIC and the PRP project we were able to connect our Science DMZ with important data sources around the country such as NERSC. It is through highly collaborative projects like this that UCSC continues to deliver on its mission to keep California on the cutting edge of scientific discoveries.”
“By setting up a campus Science DMZ, using the data-transfer node infrastructure at the NERSC facility, and using the PRP cyberinfrastructure running over CENIC’s 100 Gbps optical links, UCSC is now able to transfer data sets at faster and faster speeds. The scientific achievements in this award are enabled by high-functioning physical and human networks, both of which are essential and notable,” said Louis Fox, President and CEO of CENIC.