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:20181221T160725Z
LOCATION:D163
DTSTART;TZID=America/Chicago:20181111T110000
DTEND;TZID=America/Chicago:20181111T113000
UID:submissions.supercomputing.org_SC18_sess159_ws_indis106@linklings.com
SUMMARY:Flowzilla: A Methodology for Detecting Data Transfer Anomalies in 
 Research Networks
DESCRIPTION:Workshop\nArchitectures, Networks, Security, Workshop Reg Pass
 \n\nFlowzilla: A Methodology for Detecting Data Transfer Anomalies in Rese
 arch Networks\n\nGiannakou, Gunter, Peisert\n\nResearch networks are desig
 ned to support high volume scientific data transfers that span multiple ne
 twork links. Like any other network, research networks experience anomalie
 s. Anomalies are deviations from profiles of normality in a science networ
 k’s traffic levels. Diagnosing anomalies is critical both for network oper
 ators and scientists. In this paper, we present Flowzilla, a general frame
 work for detecting and quantifying anomalies on scientific data transfers 
 of arbitrary size. Flowzilla incorporates Random Forest Regression for pre
 dicting the size of data transfers and utilizes an adaptive threshold mech
 anism for detecting outliers. Our results demonstrate that our framework a
 chieves up to 92.5% detection accuracy. Furthermore, we are able to predic
 t data transfer sizes up to 10 weeks after training with accuracy above 90
 %.
URL:https://sc18.supercomputing.org/presentation/?id=ws_indis106&sess=sess
 159
END:VEVENT
END:VCALENDAR