BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Managing Bufferbloat in Storage Systems
DTSTART;VALUE=DATE-TIME:20201202T134500Z
DTEND;VALUE=DATE-TIME:20201202T141500Z
DTSTAMP;VALUE=DATE-TIME:20260814T190622Z
UID:indico-contribution-1311@events.chpc.ac.za
DESCRIPTION:Speakers: Carlos Maltzahn (University of California\, Santa Cr
 uz)\nScalable storage servers consist of multiple parts that communicate a
 synchronously via queues. There is usually a frontend that queues access r
 equests from storage clients and uses one or more threads to forward queue
 d requests to a backend. The backend queues these forwarded requests and b
 atches them to efficiently use storage devices it manages. Storage servers
  can have multiple kinds of backends with different design assumptions abo
 ut their underlying storage device technologies. Requests are scheduled in
  the frontend to ensure different levels of service for different classes 
 of requests. For example\, requests that are generated by data scrubbers w
 orking in the background generally have a lower priority than requests fro
 m an application. A common solution to the above problem is to move reques
 t scheduling from the frontend to the backend. For various reasons that is
  not always practical. The scope of the proposed project is to have the sc
 heduler reside in the frontend and to explore designs for backends to dyna
 mically control the admission of requests depending on continually changin
 g workloads and storage device technologies. \n\nScheduling in the fronten
 d and batching in the backend work best if there are enough requests in th
 eir respective queues. This raises the question: what is enough for the fr
 ontend and for the backend? If there are too few requests in the frontend 
 but more than enough requests in the backend\, the system might work well 
 in terms of overall throughput but might poorly enforce scheduling objecti
 ves. If there are too few requests in the backend\, then overall throughpu
 t and latency suffer no matter the scheduling objectives. If\, however\, t
 he backend has the ability to admit just enough requests from the frontend
  but not more\, throughput and latency of the backend is likely satisfacto
 ry. If there is enough work overall\, the frontend has enough requests to 
 meet scheduling objectives. How many requests are just enough for the back
 end?\n\nIn this talk I will give an overview of an ongoing research projec
 t at the UC Santa Cruz Center for Research in Open Source Software (cross.
 ucsc.edu) to reframe this question as a bufferbloat mitigation problem usi
 ng algorithms similar to the ones used for bufferbloat in networking.\n\nh
 ttps://events.chpc.ac.za/event/84/contributions/1311/
LOCATION:
URL:https://events.chpc.ac.za/event/84/contributions/1311/
END:VEVENT
END:VCALENDAR