David Lo's poster on "Heracles: Improving Resource Efficiency at Scale" (his upcoming ISCA paper) won a research award at the 2015 DiscoverEE Days at Stanford.
News feed
-
David Lo receives DiscoverEE poster award [April 2015]
-
OLDIsim & PerfKit open sourced [March 2015]
We have worked with Google to develop Perfkit, a framework for cloud performance benchmarking. One of the most interesting benchmarks in the suite is OLDIsim, that models the distributed, fan-out nature of many modern applications with tight tail latency requirements, such as Google Search and some NoSQL database applications. The original version of OLDsim was developed by David Lo.
-
Ana Klimovic receives MSR Fellowship [January 2015]
Congratulations to Ana Klimovic for receiving a PhD fellowship from Microsoft Research.
-
Workshop on Resource-Efficient Cloud Computing at ISCA 2015 [January 2015]
Prof. Christos Kozyrakis and Christina Delimitrou are organizing in collaboration with professors Lingjia Tang and Jason Mars from the University of Michigan the first workshop on Resource-Efficient Cloud Computing (REC2). The workshop will be held with ISCA-42 on June 13th in Portland, Oregon. ISCA this year is colocated with FCRC which includes several conferences in areas of systems and theory.
If you are working on areas related to improving the efficiency of large-scale systems, consider submitting to the workshop. We are aiming for a mix of early idea position papers, and invited talks from key people in the industry. The paper submission deadline is March 30th. For more details take a look at the workshop's website.
-
David Lo receives Google PhD Fellowship [September 2014]
Congratulations to David Lo for receiving a 2014 Google PhD Fellowship in Energy Efficient Computing
-
PEGASUS presented in ISCA [June 2014]
David Lo's paper on PEGASUS was presented in this year's ISCA. PEGASUS demonstrates that combining end-to-end latency monitoring with a fine-grained power management mechanism is key to tackling the energy proportionality problem for large-scale latency-critical workloads. PEGASUS was evaluated on production Google web search in a production Google search cluster and demonstrated that it can save up to 30% power without impacting latency guarantees.
-
PEGASUS in the news [June 2014]
-
MAST receives Microsoft SEIF Award [April 2014]
Our proposal on "Resource Efficient Cloud Computing" has received a 2014 Microsoft Research Award for the Software Engineering Innovation Foundation (SEIF Award).
Some press coverage on the topic by ZDnews.
-
Quasar in the news [March 2014]
Christina Delimitrou's work on Quasar has been featured in a series of articles including the Stanford Report, Stanford SoE news, The Register, The New York Times, The Stanford Daily, GigaOM, CloudPro, IBM Midsize Insider, The Whir, Slashdot, CIO Review, Financial Express, and Green Datacenter News.
-
Quasar presented in ASPLOS [March 2014]
Christina Delimitrou's paper on Quasar was presented in this year's ASPLOS. Quasar moves away from the traditional reservation-based approach in cluster management and shifts to a performance-centric approach. It leverages fast classification techniques to determine appropriate resource allocations and assignments of incoming applications and is able to both improve cluster utilization and preserve per-application QoS constraints.