BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Hermes: Clustering Users in Large-Scale E-mail Services - Christos
  Gkantsidis\, Microsoft Research Cambridge
DTSTART:20100601T103000Z
DTEND:20100601T110000Z
UID:TALK25015@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Hermes is an optimization engine for large-scale enterprise e-
 mail services. Such services could be hosted by a virtualized e-mail servi
 ce provider\, or by dedicated enterprise data centers. In both cases we ob
 serve that the pattern of e-mails between employees of an enterprise forms
  an implicit social graph. Hermes tracks this implicit social graph\, peri
 odically identifies clusters of strongly connected users within the graph\
 , and co-locates such users on the same server. Co-locating the users redu
 ces storage requirements: senders and recipients who reside on the same se
 rver can share a single copy of an e-mail. Co-location also reduces inter-
 server bandwidth usage. We evaluate Hermes using a trace of all e-mails wi
 thin a major corporation over a five month period. The e-mail service supp
 orts over 120\,000 users on 68 servers. Our evaluation shows that using He
 rmes results in storage savings of 37% and bandwidth savings of 50% compar
 ed to current approaches. The overheads are low: a single commodity server
  can run the optimization for the entire system. 
LOCATION:Small public lecture room\, Microsoft Research Ltd\, 7 J J Thomso
 n Avenue (Off Madingley Road)\, Cambridge
END:VEVENT
END:VCALENDAR