Institute of Computer Science Department of Distributed Systems Prof. Dr.-Ing. P. Tran-Gia QoEWeb: Quality of Experience and User Behaviour Modelling for Web Traffic Tobias Hoßfeld [email protected]
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Institute of Computer ScienceDepartment of Distributed Systems
Prof. Dr.-Ing. P. Tran-Gia
QoEWeb: Quality of Experience and User Behaviour Modelling for Web Traffic
Tobias Hoß[email protected]
2Tobias Hoßfeld
Partner
University of WürzburgTobias Hoßfeld, Daniel Schlosser, Thomas Zinner, Valentin Burger
Blekinge Institute of TechnologyMarkus Fiedler, Patrik Arlos, Junaid Shaikh
France Telecom SASergio Beker, Denis Collange,Frédéric Guyard, Frédérique Millo
Warsaw University of TechnologyZbigniew Kotulski, Wojciech Mazurczyk,Tomasz Ciszkowski
http://www3.informatik.uni-wuerzburg.de/research/projects/qoeweb/
3Tobias Hoßfeld
Motivation
User behavior strongly influences systemse.g. selfishness, churn, or pollution in P2P systemstime-based or volume-based models in shared systems
But, current web traffic models do not consider QoE / user behavior / impatience !
Derive QoE and user behavior model for web traffic based on active measurements in a laboratory testpassive measurements within an operator’s network
Apply model and evaluate its impact on selected exampleswireless networks with shared capacityreputation management to react before the user reacts
4Tobias Hoßfeld
Agenda
Impact of User BehaviorExample: rate control in UMTS
Active and passive measurements
QoE and User Behaviour Modelling for Web Trafficnon-linear interdependency between QoE and QoStimely behavior
Reputation Management
Work plan
5Tobias Hoßfeld
Example: Rate Control in UMTS Systems
Best-effort user and QoS user with guaranteed bandwidth
Time- and volume-based user: e.g. voice calls and FTP userImpact of user behavior on performance of system?
Transmission Powerfor QoS user? ⇒ Rate for BE user!
BS x
Time
cT
xT
Transmission Power for QoS1
QoS1
QoS2
QoS4
QoS3
BE3
BE1
BE2
Transmission Power for BE3
maxT
Constant Transmission Power for DPCCH
,x BET
,x QoST
CCHT
6Tobias Hoßfeld
A Priori Source Traffic Model of a Web User
Viewing Time V
Web Page Web Page
WWW Session WWW Session
InterSessionTime I
InlineObject
GetReq.
TCP 1TCP 2TCP 3TCP 4
GetReq.
GetReq.
Inline Object
InlineObjectGetReq. Inline Object
InlineObjectGetReq.
Inline Object
MainObject GetReq.
Web Page
Number X ofWeb Pages
GetReq.
Volume R ofGetRequest
Volume M ofMain Object
Volume O ofInline Object
Number N ofInline Objects
N
7Tobias Hoßfeld
Simulation: Web Users in Rate-Controlled UMTS
Different conclusions according to user behaviour modelvolume-based users: rate control degenerates?!time-based users: rate control works as expected?!
Important to get realistic models
volume-based users time-based users
web usersrate decreases
duration of session
increases
number of users
increases
8Tobias Hoßfeld
Basic Queueing Theory
Birth-Death-Model
Time-based users
Volume-based users
0
µ0
1
λ λ
µ1
M(1)-1 M(1) M(1)+1
λ λ λ λ
µM(1)-1 µM(1) µM(1)+1 µM(1)+2
n-1 n
λ λ
µn-1 µn
Basic queueing theory leads to same qualitative results
understanding of system behavior
will be applied in QoEWeb
9Tobias Hoßfeld
Agenda
Impact of User BehaviorExample: rate control in UMTS
Active and passive measurements
QoE and User Behaviour Modelling for Web Trafficnon-linear interdependency between QoE and QoStimely behavior
Reputation Management
Work plan
10Tobias Hoßfeld
Objectives of Measurements
Active measurementsquantification of user impatience due to bad network conditionsquantification of the decrease of satisfaction as a function of time or actionsdisturb QoS in laboratory environment user surveycan also be applied to interpret passive measurements
Passive measurementsinvestigate the statistical behavior of web trafficanalyze the correlations between the behavior of users and some network performance metrics
11Tobias Hoßfeld
Passive Measurements: Traffic Modeling
Daily behaviorTypical hours
Model of web transfers / sessionsTraffic metrics: up/down volume, type of endNetwork performance criteria: throughput, loss rate, RTTApplication level performance: response time, cancelled downloads
Type of web transfers with similar characteristicsAggregation in sessions (threshold ?)Type of web serversInfluence of the hourly variations
Model the behavior of web users, typology
12Tobias Hoßfeld
Analysis of Correlations
Correlation between traffic metrics and performance criteriaFor web transfers / sessions / users
⇒ significant performance criteria, dependence functionaccording to
the type of transfer / sessionthe type of users
13Tobias Hoßfeld
Agenda
Impact of User BehaviorExample: rate control in UMTS
Active and passive measurements
QoE and User Behaviour Modelling for Web Trafficnon-linear interdependency between QoE and QoStimely behavior
Reputation Management
Work plan
14Tobias Hoßfeld
Interdependency between QoE and QoS
Comparing iLBC and G.711 voice codecsSimilar results for both codecs regarding packet lossIQX (exponential interdepency) cannot be rejected
iLBCG.711
15Tobias Hoßfeld
Impact of Autocorrelated Delays
For different correlation factors, still exponential relationship validClear impact of correlation, i.e. timely dependencies, on QoE
16Tobias Hoßfeld
Combining Active and Passive Measurements
Viewing Time V
Web Page Web Page
WWW Session WWW Session
InterSessionTime I
InlineObject
GetReq.
TCP 1TCP 2TCP 3TCP 4
GetReq.
GetReq.
Inline Object
InlineObjectGetReq. Inline Object
InlineObjectGetReq.
Inline Object
MainObject GetReq.
Web Page
Number X ofWeb Pages
GetReq.
Volume R ofGetRequest
Volume M ofMain Object
Volume O ofInline Object
Number N ofInline Objects
N
User will •abort if QoE is too bad or•enjoy browsing and prolongs sessions for good QoE
Content and usage of web is changing• download of documents: pdf, ppt, …• video streams• services like chat or RDP
Web page may not only be provided by a single server, • but from a CDN• from different service providers (Akamai, ads server, …)
17Tobias Hoßfeld
Agenda
Impact of User BehaviorExample: rate control in UMTS
Active and passive measurements
QoE and User Behaviour Modelling for Web Trafficnon-linear interdependency between QoE and QoStimely behavior
Reputation Management
Work plan
18Tobias Hoßfeld
Reputation concept
Reputation is a proven mechanism for reflecting aggregated level of trust to network services, users, shared resources (e.g. auctioning systems, P2P networks, distributed wireless networks such as MANET; eBay, eDonkey, SecMon)Reputation management is a feedback decision process being in charge of examining the given reputation (e.g. QoE, service performance) and triggering/enforcing remedy procedures on the on-line or threshold basisKey features of reputation
present and historical measurements are weighted and reflect an its evolution and dynamicsin distributed P2P environments reputation is shared among network nodes reinforcing decision processbased on historical measurements estimates future expectations
19Tobias Hoßfeld
Reputation application in QoEWeb
Reputation buildingFor a particular Web service/Web traffic a perceived level of user satisfaction ST is expressed by QoE metrics and quantified according to the created model of user behaviourOwn experience OE of reputation is fed by ST, applyinghistorical data shaping with WMA function γFor shared reputation V service reputation SR is created with respect to credibility of recommenders IR
Reputation usage in QoEWebEvaluation of QoE metrics dynamic with respect to a particular Web services (web surfing, high throughput data, live streaming, interactive real time communication, etc)Detects deterioration of networks performance before the user perceived QoE goes down below a critical level
20Tobias Hoßfeld
Agenda
Impact of User BehaviorExample: rate control in UMTS
Active and passive measurements
QoE and User Behaviour Modelling for Web Trafficnon-linear interdependency between QoE and QoStimely behavior
Reputation Management
Work plan
21Tobias Hoßfeld
Work Plan
WP1: Measurements of web traffic
WP1.a: passiveWP1.b: active
WP3: Analysis for Business Cases
WP3.a: wirelessWP3.b: reputation
WP2: Modelling of user perception and behaviorcombine
information
apply model
future work:compare results
update information
Related Documents
