BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260522T150121Z
LOCATION:C141/143/149
DTSTART;TZID=America/Chicago:20181114T133000
DTEND;TZID=America/Chicago:20181114T140000
UID:submissions.supercomputing.org_SC18_sess177_pap229@linklings.com
SUMMARY:A Reference Architecture for Datacenter Scheduling: Design, Valida
 tion, and Experiments
DESCRIPTION:Georgios Andreadis (Delft University of Technology, Vrije Univ
 ersity Amsterdam); Laurens Versluis (Vrije University Amsterdam); Fabian M
 astenbroek (Delft University of Technology); and Alexandru Iosup (Vrije Un
 iversity Amsterdam, Delft University of Technology)\n\nDatacenters act as 
 cloud-infrastructure to stakeholders across industry, government, and acad
 emia. To meet growing demand yet operate efficiently, datacenter operators
  employ increasingly more sophisticated scheduling systems, mechanisms, an
 d policies. Although many scheduling techniques already exist, relatively 
 little research has gone into the abstraction of the scheduling process it
 self, hampering design, tuning, and comparison of existing techniques. In 
 this work, we propose a reference architecture for datacenter schedulers. 
 The architecture follows five design principles: components with clearly d
 istinct responsibilities, grouping of related components where possible, s
 eparation of mechanism from policy, scheduling as complex workflow, and hi
 erarchical multi-scheduler structure. To demonstrate the validity of the r
 eference architecture, we map to it state-of-the-art datacenter schedulers
 . We find scheduler-stages are commonly underspecified in peer-reviewed pu
 blications. Through trace-based simulation and real-world experiments, we 
 show underspecification of scheduler-stages can lead to significant variat
 ions in performance.\n\nTag: Clouds and Distributed Computing, Resource Ma
 nagement, Scheduling\n\nRegistration Category: Tech Program Reg Pass\n\nSe
 ssion Chair: Shantenu Jha (Rutgers University; PPPL, Princeton University)
 \n\n
END:VEVENT
END:VCALENDAR