BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Asia/Calcutta
BEGIN:STANDARD
DTSTART:19451014T230000
TZOFFSETFROM:+0630
TZOFFSETTO:+0530
TZNAME:IST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20130118T073726Z
UID:9F79340D-166E-11E8-9184-0050568D7F66
DTSTART;TZID=Asia/Calcutta:20120801T160000
DTEND;TZID=Asia/Calcutta:20120801T170000
DESCRIPTION:*Title*: Dynamic Monitoring and Optimization in Communication N
 etworks *Speaker*: Ketan Rajawat *Abstract*: Communication networks have e
 volved from specialized\, research- and military-oriented transmission sys
 tems to large-scale and highly complex interconnections of intelligent dev
 ices. Effective operation of such large-scale networks hinges upon real-ti
 me allocation of network resources that match the user demands. My PhD the
 sis contributes towards several key problems encountered in both\, monitor
 ing and resource allocation in networks. My talk will focus on the problem
  of mapping the network state from incomplete sets of measurements\, and w
 ill touch upon two application domains. The first part of my talk consider
 s the problem of tracking and predicting end-to-end network delay. The foc
 us here is on tools that can take advantage of the spatio-temporal correla
 tions exhibited by the end-to-end delay. A dynamic network kriging approac
 h is introduced that not only allows efficient tracking and prediction\, b
 ut also online selection of measurement locations. The second part of my t
 alk considers traffic volume measurements\, which are often replete with m
 issing data. A widely studied problem in this context is that of estimatin
 g origin-destination traffic matrices\, with most techniques relying on st
 ationarity assumptions regarding traffic volumes. A novel approach utilizi
 ng dictionary learning for imputing missing link traffic volumes is introd
 uced\, and is shown to efficiently handle a much wider class of non-statio
 nary traffic patterns. In the last part of the talk\, I will briefly descr
 ibe my work in resource allocation for ad hoc and sensor networks. Here we
  leverage the idea of network coding to design cross-layer protocols that 
 are both throughput optimal and energy efficient. *Bio*: Ketan Rajawat rec
 eived his B. Tech and M. Tech degrees in Electrical Engineering from India
 n Institute of Technology Kanpur\, in 2007\; and his Ph.D. degree in Elect
 rical and Computer Engineering from the U. of Minnesota\, in 2012. His res
 earch interests lie in the areas of SP\, communication networks\, and wire
 less communications. His current research focuses on network optimization 
 and monitoring.\n\nCo-sponsored by: Dr. Kumar Vaibhav Srivastava\n\nKanpur
 \, Uttar Pradesh\, India
LOCATION:Kanpur\, Uttar Pradesh\, India
ORGANIZER:kvs@iitk.ac.in
SEQUENCE:0
SUMMARY:[Legacy Report] Dynamic Monitoring and Optimization in Communicatio
 n Networks
URL;VALUE=URI:https://events.vtools.ieee.org/m/165371
X-ALT-DESC:Description: <br />*Title*: Dynamic Monitoring and Optimization 
 in Communication Networks\n\n*Speaker*: Ketan Rajawat\n\n*Abstract*: Commu
 nication networks have evolved from specialized\, \nresearch- and military
 -oriented transmission systems to large-scale and \nhighly complex interco
 nnections of intelligent devices. Effective \noperation of such large-scal
 e networks hinges upon real-time allocation \nof network resources that ma
 tch the user demands. My PhD thesis \ncontributes towards several key prob
 lems encountered in both\, monitoring \nand resource allocation in network
 s.\n\nMy talk will focus on the problem of mapping the network state from 
 \nincomplete sets of measurements\, and will touch upon two application \n
 domains. The first part of my talk considers the problem of tracking and \
 npredicting end-to-end network delay. The focus here is on tools that can 
 \ntake advantage of the spatio-temporal correlations exhibited by the \nen
 d-to-end delay. A dynamic network kriging approach is introduced that \nno
 t only allows efficient tracking and prediction\, but also online \nselect
 ion of measurement locations. The second part of my talk considers \ntraff
 ic volume measurements\, which are often replete with missing data. \nA wi
 dely studied problem in this context is that of estimating \norigin-destin
 ation traffic matrices\, with most techniques relying on \nstationarity as
 sumptions regarding traffic volumes. A novel approach \nutilizing dictiona
 ry learning for imputing missing link traffic volumes \nis introduced\, an
 d is shown to efficiently handle a much wider class of \nnon-stationary tr
 affic patterns.\n\nIn the last part of the talk\, I will briefly describe 
 my work in \nresource allocation for ad hoc and sensor networks. Here we l
 everage the \nidea of network coding to design cross-layer protocols that 
 are both \nthroughput optimal and energy efficient.\n\n*Bio*: Ketan Rajawa
 t received his B. Tech and M. Tech degrees in \nElectrical Engineering fro
 m Indian Institute of Technology Kanpur\, in \n2007\; and his Ph.D. degree
  in Electrical and Computer Engineering from \nthe U. of Minnesota\, in 20
 12. His research interests lie in the areas of \nSP\, communication networ
 ks\, and wireless communications. His current \nresearch focuses on networ
 k optimization and monitoring.
END:VEVENT
END:VCALENDAR