BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Emina Soljanin (Rutgers University)
DTSTART:20210804T160000Z
DTEND:20210804T170000Z
DTSTAMP:20260416T224402Z
UID:CCM2021/1
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/CCM2021/1/">
 Codes\, Graphs\, and Hyperplanes in Data Access Service</a>\nby Emina Solj
 anin (Rutgers University) as part of Carleton Combinatorics Meeting 2021\n
 \n\nAbstract\nDistributed computing systems strive to maximize the number 
 of concurrent data access requests they can support with fixed resources. 
 Replicating data objects according to their relative popularity and access
  volume helps achieve this goal. However\, these quantities are often unpr
 edictable. Erasure-coding has emerged as an efficient and robust form of r
 edundant storage. In erasure-coded models\, data objects are elements of a
  finite field\, and each node in the system stores one or more linear comb
 inations of data objects. This talk asks 1) which data access rates an era
 sure-coded system can support and 2) which codes can support a specified r
 egion of access rates. We will address these questions by casting them int
 o some known and some new combinatorial optimization problems on graphs. W
 e will explain connections with batch codes. This talk will also describe 
 how\, instead of a combinatorial\, one can adopt a geometric approach to t
 he problem.\n
LOCATION:https://researchseminars.org/talk/CCM2021/1/
END:VEVENT
END:VCALENDAR