Clustered File Systems for High Performance Computing and Big Data by Doug Oucharek Manager, Intel Corporation (26 Nov. 2013)

DougOucharek26Nov2013Doug Oucharek

Manager of the Lustre Core Development group, Intel Corporation

 Clustered File Systems for High Performance Computing and Big Data

Time & Date: 5pm-6pm, Tuesday November 26th, 2013

Location: E 103, Okanagan College, Kelowna Campus, 1000 KLO Rd., Kelowna, BC V1Y 4X8

Registration is opened:

Talk Abstract: The world of HIgh Performance Computing (HPC) is one of vast numbers: tens of thousands of compute nodes clustered together, petabytes of shared storage, and kilometres of network cables.  The demands HPC places upon storage systems are the most extreme of any other computing discipline.  In this talk, I will be outlining the specific requirements of HPC storage systems, and delve into the architecture of “Lustre”: an open source parallel distributed file system used by the majority of the top 100 supercomputers.  I will also be talking about the future of HPC file systems and how Big Data is adopting HPC file systems to improve performance.

Speaker Biography: Doug Oucharek is the manager of the Lustre Core Development group at Intel.  He was pulled into Intel as part of an acquisition of a startup supporting Lustre called “Whamcloud”.  Prior to working for Whamcloud, he spent over 25 years working on various networking products at several companies including Nortel, Motorola, IBM, HP, and Broadcom.  Doug works and lives in Naramata, BC, Canada.