BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Pelican: A Building Block for Exascale Cold Data Storage - Austin 
 Donnelly (Microsoft Research\, Cambridge)
DTSTART:20160121T150000Z
DTEND:20160121T160000Z
UID:TALK62876@talks.cam.ac.uk
CONTACT:Liang Wang
DESCRIPTION:Pelican is a rack-scale design for cheap storage of data which
  is rarely accessed: cold data. It uses spun-down hard drives to maximise 
 density and reduce costs. A Pelican rack supplies only enough resources (p
 ower\, cooling\, bandwidth) to support the cold data workloads we target\,
  significantly reducing Pelican's total cost of ownership compared to trad
 itional disk-based systems provisioned for peak performance. The Pelican s
 torage stack manages the limited resources\, and their constraints. We des
 cribe the data layout and IO scheduling algorithms which ensures these con
 straints are not violated\, while making best use of the available resourc
 es. We evaluate Pelican both in simulation and with a full rack\, and show
  that Pelican performs well: delivering both high throughput and acceptabl
 e latency.\n\nBio: Austin Donnelly is a Principal Research Software Develo
 pment Engineer (RSDE) in the Systems group at Microsoft Research\, Cambrid
 ge. He obtained his B.A. in Computer Science from the University of Cambri
 dge in 1996\, and went on to complete his Ph.D. there in 2002.  He has wor
 ked on storage technologies for around 20 years\, from writing an IDE driv
 er for Nemesis (a research OS)\, to research on the write-offloading techn
 ique used in DiskEnergy to save power.  He is currently working on the Pel
 ican system for cold storage in datacentre environments.
LOCATION:FW26\, Computer Laboratory\, William Gates Building
END:VEVENT
END:VCALENDAR