InTheLoop | 10.07.2013
CS Staff Contribute to NSF Cybersecurity Summit
Last week, the National Science Foundation sponsored the2013 NSF Cybersecurity Summit for Cyberinfrastructure and Large Facilities in Arlington, Va., and staff from NERSC, CRD and ESnet participated. The theme of this year’s meeting was “Designing Cybersecurity Programs in Support of Science.”
Fitting into that theme was Vern Paxson, who holds a visiting faculty appointment in CRD’s Advanced Computing for Science Department gave and gave the Oct. 2 keynote talk on community building for cybersecurity. As a lab employee in the 1990s, Paxson developed Bro, an intrusion detection system long used by the lab for cybersecurity. Bro is now being commercialized and adopted by both public and private sector organizations. As part of the meeting, a half-day tutorial on Bro was presented on Sept. 30.
“The real star of the meeting was Bro, which was discussed - and is being deployed – extensively,” said ESnet Director Greg Bell, who led a panel discussion on the differences, similarities and relationships between NSF centers and other research institutions. Also attending the invitation-only meeting were Michael Sinatra of ESnet and Scott Campbell of NERSC.
CRD’s Wehner, Donofrio Quoted in Article on Improving Climate Computing
Michael Wehner of the Scientific Computing Group and Dave Donofrio of the Future Technologies Group were interviewed and quoted in an article about developing more powerful computers developed specifically for climate modeling. For his Oct. 4 article in ClimateWire, reporter Umair Irfan spent a day at Berkeley Lab interviewing researchers about computing advances to improve the study of climate change.
Current climate models are limited both by the amount of data needed to study problems at finer scales and the hardware itself. “As a result, scientists are exploring the principle of co-design to improve climate modeling, developing computing hardware and software at the same time with the same goals in mind,” Irfan wrote. “With customized machines, researchers expect they can do more with less.”
Wehner, who is a lead author of the climate projections chapter in the latest IPCC assessment report, told Irfan that researchers have recently managed to resolve climate models down to 25 kilometers. “"What we get from that is much better extreme weather. The storms look like real storms, whereas at 100 kilometers, they don't," Wehner said. "What we're getting there is some enhanced realism, but we're not all the way [there]."
Donofrio was also quoted about the lab’s GreenFlash project aimed at using smaller, more energy efficient processors to develop specialized supercomputers for climate research. »Download the story (PDF | 620KB).
This Week’s Computing Sciences Seminars
Combinatorial and statistical algorithms for cellular informatics
Monday, October 7, 10 – 11 a.m., Bldg. 50F, Room 1647
Ariful Azad, Department of Computer Science, Purdue University
Abstract: In this talk I will present my ongoing research on parallel algorithms to compute match-ings in graphs and their applications to analyze high-dimensional, cell-level proteomicsdata. At first I will discuss matching algorithms in a general context and present several classes of parallel algorithms to compute maximum matching in large graphs on shared-memory achines. I will then discuss the application of matching and other combinatorial algorithms to solve “Big Data" problems from single-cell proteomics study.
Fast Scientific Data Analysis with Statistical Metadata
Monday, October 7, 11 a.m. – 12 noon, Bldg. 50B, Room 2222
Yong Chen, Computer Science Department, Texas Tech University
Scientific datasets and libraries, such as HDF5, ADIOS, and NetCDF, have been used widely in many data intensive scientific simulations and applications. These libraries have their special file formats and I/O functions to provide efficient access to large datasets. Recent studies have started to utilize indexing, sub-setting, and data reorganization to manage the increasingly large datasets. In this talk, I will present our recent study of Fast Analysis with Statistical Metadata (FASM) intending to boost the data analysis performance via data sub-setting and integrating a small amount of statistics into the original datasets. The added statistical information illustrates the data shape and provides knowledge of the data distribution; therefore the original I/O libraries can utilize these statistical metadata to perform fast queries and analyses. We will also introduce segmented analysis, pre-analysis, and decoupled execution paradigm concepts and ideas for reducing data movements and to speed up data analysis for scientific discoveries. These concepts and ideas can potentially lead to new data analytics methodologies and can have an impact on big data analysis.
Link of the Week: Getting Dialed in on the Nobel Prizes
This week, Prof. Staffan Normark, permanent secretary of the Royal Swedish Academy of Sciences, will make life-changing calls informing two scientists informing them that they have won the Nobel Prize for Chemistry (on Tuesday) and Physics (on Wednesday). An article in the Economist's Intelligent Life goes behind the scenes with Prof. Normark. Read more at http://moreintelligentlife.com/content/ideas/anonymous/nobel-calling