InTheLoop | 05.12.2014
Media Turn to CRD's Wehner to Explain Climate Change Impacts
In an interview with KRON-TV Channel 4 last week, the Computational Research Division's Michael Wehner told Bay Area TV newswatchers to expect precipitation patterns to shift as climate change progresses. (The broadcast is archived in two parts: »Part One »Part Two.) Stories on the release of the National Climate Assessment in the San Francisco Business Times and Daily Cal story also noted Wehner's role as a lead author of the report released by the White House last week.
ESnet Staff Speak at NSF Workshop on Campus Network Infrastructure for Science
On April 30-May 1, the National Science Foundation hosted a workshop for recipients of 2012 and 2013 NSF awards from the Campus Cyberinfrastructure-Network Infrastructure and Engineering (CC-NIE) initiative. The initiative aims to help universities develop and deploy network infrastructures designed to speed the flow of scientific datasets in and out of research institutions. A key component of the design is the Science DMZ, a model that addresses common network performance problems encountered at research institutions by creating an environment that is tailored to the needs of high performance science applications, including high-volume bulk data transfer, remote experiment control, and data visualization. The following ESnet staff members participated in the workshop program:
- Eli Dart gave a joint talk on the Science DMz architecture, including an overview, discussion of security and a look at what’s next.
- Jason Zurawski was one of two speakers talking about integrating new technologies into science applications.
- Zurawski also spoke in a session on deployments of perfSONAR, software co-developed by ESnet to identify network performance bottlenecks.
- Lauren Rotman gave an invited talk on ESnet’s program for gathering science and networking requirements from ESnet users.
More than 100 people, mostly from universities, attended the workshop in Arlington, Va.
Unat Organizes Workshop to Address Data Locality Challenges in Programming
Didem Unat, with the Computational Research Division, was one of the organizers of a workshop held April 28 & 29 in Lugano, Switzerland to address the emerging challenges of programming for changing high-performance computing (HPC) data architectures. The goal of the Workshop on Programming Abstractions for Data Locality (PADAL) was to identify common themes and standardize concepts for locality-preserving abstractions for exsacale programming models. Participants are compiling a report to share their findings with the rest of the HPC community. »Learn more
Dosanjh, Antypas Take InsideHPC Under the Hood of NERSC's Next-Gen 'Cori' System
In a May 9 InsideHPC interview. Sudip Dosanjh,director of NERSC, and Katie Antypas, NERSC department head, discussed the newly announced "Cori" supercomputer. NERSC and Cray recently inked the contract for the next-gen system expected to arrive after the center moves into the Computational Research and Theory facility, its new home now under construction on the main. »Read more
Mark Your Calendars: NERSC Nobel Lectures Start Next Week
Just a reminder that the lunchtime series of Nobel-winning NERSC users starts next Tuesday, May 20. For a reminder of the lineup, days and locations »Visit the NERSC site.
This Week's CS Seminar
Cloud Datacenter Resource Management - a bird's eye view of Umeå University's Cloud research
Tuesday, May 13, 2014, 2:00pm - 3:00pm, Bldg. 50B, Room 4205
P-O Östberg, Umeå University, Sweden
Taking a holistic approach to cloud resource management, Umeå University's distributed systems group aims to transform today's static and energy consuming cloud data centers into self-managed, dynamic, and dependable automated infrastructures. This presentation will give an overview of UmU's research efforts with glimpses into selected projects ranging from low-level live migration algorithms to high-level holistic optimization of data center resource utilization. Selected research topics include, e.g., addressing fundamental self-management challenges that arise from stochastic variations in capacity requirements, resource availability, and operation costs, or optimization of energy efficiency in cloud data centers subject to differentiated and variable application quality of service. For more information about cloud resource management research at Umeå University, Sweden, please visit www.cloudresearch.org.