Post on 30-May-2018
8/9/2019 Internet QoS UMKC Frost
1/42
University of KansasA KTEC Center of Excellence 1
Soshant Bali*, Yasong Jin**, Victor S. Frost* andTyrone Duncan**
Information and Telecommunication Technology Center*Electrical Engineering & Computer Science
**Department of Mathematicsfrost@eecs.ku.edu, 785-864-4833
A New Perspective on Internet
Quality of Service: Measurement andPredictions
8/9/2019 Internet QoS UMKC Frost
2/42
8/9/2019 Internet QoS UMKC Frost
3/42
University of KansasA KTEC Center of Excellence 3
Outline
Develop end-to-end measurement techniques Develop prediction methodologies for fBM
traffic
A Few Words about our Graduate andResearch Programs at EECS@KU
8/9/2019 Internet QoS UMKC Frost
4/42
University of KansasA KTEC Center of Excellence 4
Premise
Voice networks had a very understandable QoS metric-Blocking Internet QoS metrics must correlate to end-user experience. Metrics such as delay and loss may have little direct meaning
to the end-user because knowledge of specific coding and/oradaptive techniques is required to translate delay and loss tothe user-perceived performance.
Detecting observable impairments must be independent ofcoding, adaptive playout or packet loss concealment techniquesemployed by the multimedia applications.
Time between impairments and their duration are metrics thatare easily understandable by network user.
This research developed methods to detect these impairmentevents using end-to-end measurements.
8/9/2019 Internet QoS UMKC Frost
5/42
University of KansasA KTEC Center of Excellence 5
Network states
Noticeable impairments for Real-time multi-media (RTM) services occur when the end-to-end connection is in one or more of thefollowing states: Burst loss, High random loss,
Disconnected,
High Delay.
Two other connection states are defined: Congested, Route change.
8/9/2019 Internet QoS UMKC Frost
6/42
University of KansasA KTEC Center of Excellence 6
Background End-to-end argument
end nodes: most functions implemented here including applicationspecific functions core: important forwarding and routing functions are implemented
here; not burdened by application specific functions, e.g., reliabledelivery
Anomalous events failures: fiber cuts, power failures etc. congestion cause user-perceived impairments
Inferring anomalous events from end-to-endobservations core nodes implement simple functions; do not inform end nodes of
anomalous events need to infer anomalous events from end-to-end observations
Several benefits if anomalous events are accuratelyinferred
8/9/2019 Internet QoS UMKC Frost
7/42
University of KansasA KTEC Center of Excellence 7
Significance
A new QoS metric for RTM applications ISPs can use impairments metric in service level agreements (SLAs)
Fault diagnosis tools for ISPs an alternative to traceroute for detecting layer 3 route changes
method for detecting layer 2 failures
Routing for overlay / content delivery networks Increasing TCP throughput
Confidence interval for minimum RTT estimate(byproduct)
8/9/2019 Internet QoS UMKC Frost
8/42
University of KansasA KTEC Center of Excellence 8
Goal
Given a set of active end-to-end networkmeasurements determine the networkstate and the temporal characteristics ofimpairment events
Network
Round Trip Time
Packet Loss Rate
Traceroute
Time-to-live
Network
State
Impairment Events:
-Frequency
-Duration
8/9/2019 Internet QoS UMKC Frost
9/42
University of KansasA KTEC Center of Excellence 9
Goal
8/9/2019 Internet QoS UMKC Frost
10/42
University of KansasA KTEC Center of Excellence 10
Route Change
Motivation Route changes can cause user perceived impairments
Need to divide observations into homogenous regions
Layer 3 route changes
TTL Traceroute
Not all route changes result in TTL change
Not all routers respond to ICMP massages for traceroute
Layer 2 route changes are not visible end-to-end
8/9/2019 Internet QoS UMKC Frost
11/42
University of KansasA KTEC Center of Excellence 11
Route change state
RTT based route change detection TTL change: not all route changes result in TTL change traceroute change: inefficient, not all routers respond to ICMP massages for traceroute both layer 2 and layer 3 route changes can be detected using RTT based route change
detection
in figure below, minimum RTT changed but traceroute and TTL field of IPheader did not change; layer 2 route change
8/9/2019 Internet QoS UMKC Frost
12/42
University of KansasA KTEC Center of Excellence 12
Route Change
Layer 2 Route ChangeIf
the time between changes > T
and the RTT difference acrossthe route change > RTT
and variation in RTT
8/9/2019 Internet QoS UMKC Frost
13/42
University of KansasA KTEC Center of Excellence 13
Congested State
Observed from M/M/1Queues
is an indicator of congestion
The end-to-end flow is in theCongested sate if:
Where
= Ave waiting time
= Packet loss rate
8/9/2019 Internet QoS UMKC Frost
14/42
University of KansasA KTEC Center of Excellence 14
Congested State
RTTs and a congestion event detected using the discussed procedure
planetlab2.ashburn.equinix.planet-lab.org and planetlab1.comet.columbia.edu, 2/04
8/9/2019 Internet QoS UMKC Frost
15/42
University of KansasA KTEC Center of Excellence 15
Delay Impairment State
Given the RTT data, anestimate is made of theminimum playout delaybuffer size that isneeded to avoid
excessive packet losses. If minimum playout
delay > Dplayout then a delayimpairment has
occurred.
Estimated one-way delays and
minimum playout delay
planetlab2.ashburn.equinix.planet-lab.org
and planetlab1.comet.columbia.edu
Feb, 2004
8/9/2019 Internet QoS UMKC Frost
16/42
University of KansasA KTEC Center of Excellence 16
Other Networks States
Disconnected state Period of consecutive packet losses > sec
Burst loss state sec < Period of consecutive packet losses < sec
High Random Loss State Insure enough observed losses, e.g., N, for valid loss
probability estimate, RoT N > 10
Observe N losses, if number of packets between the firstand Nth loss < NL then network in high lose state
8/9/2019 Internet QoS UMKC Frost
17/42
University of KansasA KTEC Center of Excellence 17
Measurement data
8/9/2019 Internet QoS UMKC Frost
18/42
University of KansasA KTEC Center of Excellence 18
Congestion Eventsobserved over a period of one week (DC1)
8/9/2019 Internet QoS UMKC Frost
19/42
University of KansasA KTEC Center of Excellence 19
Statistics of user-perceived impairments
8/9/2019 Internet QoS UMKC Frost
20/42
University of KansasA KTEC Center of Excellence 20
Other observations
Layer 2 route change 96 events were manually classified as layer 2 route changes
~71.8% layer 2 route changes were detected by thealgorithm
~4% of the detected events were false positives.
~8% of all layer 3 route changes werepreceded by burst or disconnect loss events.
8/9/2019 Internet QoS UMKC Frost
21/42
University of KansasA KTEC Center of Excellence 21
Summary of measurement results
mean time between impairments: from 3.52hrs to 268hrs
mean duration of impairments: from 4.4mins to 92.5mins on 2 paths congestion for 6-8 hrs during day (weekdays)
burst loss, high random loss and high delay events were observed whenconnection was in congested state
mean time between burst loss events that occurred during congestion = 14min, mean duration = 22.64 sec
mean time between layer 3 route changes = 7.23 hrs 18% of all layer 3 route changes 1 sec apart, 15% 2 sec apart, 80% less
than 45 mins apart 8% of all layer 3 route changes were preceeded by burst or disconnect
loss events mean duration of burst loss events that precede layer 3 route changes =
113.5 sec
mean time between layer 2 route changes = 58.22 hrs
none of the layer 2 route changes were preceded by burst loss events
8/9/2019 Internet QoS UMKC Frost
22/42
University of KansasA KTEC Center of Excellence 22
Experimental Conclusions Developed procedures to detect impairment
states for RTM services using end-to-endmeasurements.
Developed techniques to detect layer tworoute changes and congestion
The developed techniques consider multiplemetrics at the same time to infer thepresence of user perceived impairments.
Details in Characterizing User-perceived Impairment Events UsingEnd-to-End Measurements, Soshant Bali, Yasong Jin, V. S. Frost and T. Duncan,International Journal of Communication Systems.
8/9/2019 Internet QoS UMKC Frost
23/42
University of KansasA KTEC Center of Excellence 23
Predicting Properties of Congestion Events
QueueSize
inBits
8/9/2019 Internet QoS UMKC Frost
24/42
University of KansasA KTEC Center of Excellence 24
Predicting Properties of Congestion Events
QueueSize
inBits
8/9/2019 Internet QoS UMKC Frost
25/42
University of KansasA KTEC Center of Excellence 25
Predicting Properties of Congestion Events
Traffic Model fractional Brownian motion (fBm)
Qo(t) = Queue length at t
=Service rate
m=average input rate
a=variance of the input rate
BH(t)=standard fBm with parameter H
c=scaled surplus rate
8/9/2019 Internet QoS UMKC Frost
26/42
University of KansasA KTEC Center of Excellence 26
Sojourn Time
8/9/2019 Internet QoS UMKC Frost
27/42
University of KansasA KTEC Center of Excellence 27
Inter congestion event time
8/9/2019 Internet QoS UMKC Frost
28/42
University of KansasA KTEC Center of Excellence 28
Congestion duration
8/9/2019 Internet QoS UMKC Frost
29/42
University of KansasA KTEC Center of Excellence 29
Amplitude
8/9/2019 Internet QoS UMKC Frost
30/42
University of KansasA KTEC Center of Excellence 30
Conclusions
Developed methods to measure impairmentsusing end-to-end measurements Developed techniques to predict several
properties of congestion events for fBMtraffic: Rate, Duration, Amplitude For details see: Predicting Properties of Congestion Events
for a Queueing System with fBM Traffic, Yasong Jin,Soshant Bali. Tyrone Duncan, Victor S. Frost, acceptedpending revisions for the IEEE Transactions on Networking.
A F W d b t G d t
8/9/2019 Internet QoS UMKC Frost
31/42
University of KansasA KTEC Center of Excellence 31
A Few Words about our GraduateProgram at EECS@KU
37 faculty 4 Fellows of the IEEE Ex-Program Managers from DARPA, NSF, NASA 10 new faculty in the past 3 years Currently recruiting one more faculty member
MS degrees in EE, CoE, CS
150 MS students Ph.D. degrees in EE, CS
75 Ph.D. students
Two major research labs: ITTC and CReSIS Research volume of over $20 million, with research expenditures of $5.5
million in 2005 >50% of our graduate students are supported (over 140 in F05)
Almost all our Ph.D. students are supported
8/9/2019 Internet QoS UMKC Frost
32/42
University of KansasA KTEC Center of Excellence 32
EECS Research Space
Wh t S f O R t G d t
8/9/2019 Internet QoS UMKC Frost
33/42
University of KansasA KTEC Center of Excellence 33
What Some of Our Recent GraduatesAre Doing Now
Cory Beard (PhD EE 1999) Associate Professor UMKC
Jennifer Leopold (PhD CS 2000) - Professor of CS at Missouri, Rolla Amit Kulkarni (PhD CS 2000) - GE Global Research Center
Daniel Cliburn (PhD CS 2001) - Professor of CS at Hanover College
Nathan Goodman (PhD EE 2002) - Professor of ECE at the University of Arizona
Cindy Kong (PhD CS 2004) - Intel Corp.
Wesam Alanqar (PhD EE 2005) - Sprint Corp.
Jungwoo Ryoo (PhD CS 2005) - Professor at Arizona State University
David Janzen (PhD CS 2006) - Professor at Cal Poly, San Louis Obispo
8/9/2019 Internet QoS UMKC Frost
34/42
University of KansasA KTEC Center of Excellence 34
Ph.D. Focus Areas
Communication Systems and Networking Computer Systems Design
Interactive Intelligent Systems
Bioinformatics
Radar Systems and Remote Sensing
8/9/2019 Internet QoS UMKC Frost
35/42
University of KansasA KTEC Center of Excellence 35
Communication Systems and Networking
Advancing knowledge of systemsinterconnected via radio and othertechnologies
New methodologies to determine the
performance and protection of Internet-based systems
Theory and technologies that enable thedelivery of reliable information in support of
end-user applications independent of theaccess technology
8/9/2019 Internet QoS UMKC Frost
36/42
University of KansasA KTEC Center of Excellence 36
Computer Systems Design
Design of computing systems, ranging from small,embedded elements to large, distributed computingenvironments
All aspects of the system life cycle, includingspecification, verification, implementation and
synthesis, and testing and evaluation of bothhardware and software system components
Principle application area of embedded and real-timesystems with special emphasis on the interaction
between hardware and software system components
8/9/2019 Internet QoS UMKC Frost
37/42
University of KansasA KTEC Center of Excellence 37
Interactive Intelligent Systems
Create intelligent and interactive systems withsufficient intelligence to help humans accomplishimportant tasks
Multi-modal interfaces to respond intelligently touser requests, process and present large quantities
of information in many forms, and to perform taskswith minimal supervision Artificial intelligence, intelligent agents, information
retrieval, data mining, human-computer interaction,modeling, visualization, multimedia systems, androbotics
8/9/2019 Internet QoS UMKC Frost
38/42
University of KansasA KTEC Center of Excellence 38
Bioinformatics
Information technology to process, analyze,and present biological data in new,meaningful, and efficient ways
Knowledge discovery and data mining andanalysis as they relate to life sciences
Making key advances in bioinformaticsmethods and tools for genomics andproteomics data analysis and other life-sciences-related problems
8/9/2019 Internet QoS UMKC Frost
39/42
University of KansasA KTEC Center of Excellence 39
Radar Systems and Remote Sensing
Radars, microwaves, communications, andremote sensing technologies New ways to use electromagnetic waves in
the remote sensing of the land (surface andsubsurface), sea, polar ice, and theatmosphere
Developing new remote sensing sensors(primarily radar), and new methods forsolving electromagnetic problems
8/9/2019 Internet QoS UMKC Frost
40/42
University of KansasA KTEC Center of Excellence 40
FastTrack Ph.D.
Enter the Ph.D. program directly from theB.S.
Finish in 5 years
42 course credit hours past B.S.
Possible schedule:Semester 1 3 courses
Semester 2 3 courses
Semester 3 2-3 courses + research
Semesters 4-10 0-2 courses + research
8/9/2019 Internet QoS UMKC Frost
41/42
University of KansasA KTEC Center of Excellence 41
Deadlines
The application deadline is March 1st, but forfull consideration for fellowships andresearch/teaching assistantships,applications should be received by January
1st. For more details about the applicationprocess please see our graduate admissionspage.
8/9/2019 Internet QoS UMKC Frost
42/42
University of KansasA KTEC Center of Excellence 42
Websites
www.ittc.ku.edu www.cresis.ku.edu
www.eecs.ku.edu
www.ku.edu