BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Scheduling and large deviations - Zwart\, AP (CWI)
DTSTART:20100322T140000Z
DTEND:20100322T150000Z
UID:TALK23838@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:We investigate the impact of a scheduling discipline in preven
 ting large response times in a queueing setting. In particular\, for a GI/
 GI/1 queue it is well known that\, in a large deviations setting\, FIFO is
  optimal for light-tailed service times\, and service disciplines such as 
 PS (Processor Sharing) and SRPT (Shortest Remaining Processing Time) are o
 ptimal for heavy-tailed service times. It is also known that PS and SRPT d
 o not perform well for light-tailed service times\, while FIFO does not pe
 rform well for heavy-tailed service times. In this talk\, we review these 
 results\, and answer the natural question whether it is possible to constr
 uct a single scheduling discipline that is competitive with FIFO for light
 -tailed service times and competitive with SRPT for heavy tailed service t
 imes. We also investigate how more robust scheduling disciplines can be de
 signed by exploiting partial information on the distributions\, such as th
 e system load. 
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR