BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20200129T163603Z
LOCATION:501
DTSTART;TZID=America/Denver:20191117T120500
DTEND;TZID=America/Denver:20191117T123000
UID:submissions.supercomputing.org_SC19_sess103_ws_indis107@linklings.com
SUMMARY:Estimation of RTT and Loss Rate of Wide-Area Connections Using MPI
  Measurements
DESCRIPTION:Workshop\n\nEstimation of RTT and Loss Rate of Wide-Area Conne
 ctions Using MPI Measurements\n\nRao, Imam, Kettimuthu, Liu, Foster\n\nSci
 entific computations are expected to be increasingly distributed across wi
 de-area networks, and Message Passing Interface (MPI) has been shown to sc
 ale to support their communications over long distances. The execution tim
 es of MPI basic operations over long distance connections reflect the conn
 ection length and losses, which should be accounted for by the application
 s, for example, by rolling back to a single site under high network loss c
 onditions. We utilize execution time measurements of MPI_Sendrecv operatio
 ns collected over emulated 10Gbps connections with 0-366ms round-trip time
 s, wherein the longest connection spans the globe, under up to 20\% period
 ic losses. We describe five machine leaning methods to estimate the connec
 tion RTT and loss rate from these MPI execution times. They provide dispar
 ate, namely, linear and non-linear, and smooth and non-smooth, estimators 
 of RTT and loss rate. Our results show that accurate estimates can be gene
 rated at low loss rates but become inaccurate at loss rates 10% and higher
 . Overall, these results constitute a case study of the strengths and limi
 tations of machine learning methods in inferring network-level parameters 
 using application-level measurements.\n\nTag: Workshop Reg Pass, Big Data,
  Data Analytics, Datacenter, Networks, Software-defined networking\n\nRegi
 stration Category: Workshop Reg Pass, Big Data, Data Analytics, Datacenter
 , Networks, Software-defined networking
URL:https://sc19.supercomputing.org/presentation/?id=ws_indis107&sess=sess
 103
END:VEVENT
END:VCALENDAR