Performance Analysis Proposal



Similar documents
PERFORMANCE ANALYSIS OF VOIP TRAFFIC OVER INTEGRATING WIRELESS LAN AND WAN USING DIFFERENT CODECS

Planning Networks for VOIP. An Introduction

12 Quality of Service (QoS)

Evaluating Data Networks for Voice Readiness

A Performance Study of VoIP Applications: MSN vs. Skype

Performance Analysis of VoIP Codecs over BE WiMAX Network

Planning an IP Network for Voice and Data

Delivering reliable VoIP Services

APTA TransiTech Conference Communications: Vendor Perspective (TT) Phoenix, Arizona, Tuesday, VoIP Solution (101)

Performance Analysis of VoIP Codecs over Wi-Fi and WiMAX Networks

VoIP Bandwidth Considerations - design decisions

Chapter 4. VoIP Metric based Traffic Engineering to Support the Service Quality over the Internet (Inter-domain IP network)

VoIP over Wireless Opportunities and Challenges

Traffic Characterization and Perceptual Quality Assessment for VoIP at Pakistan Internet Exchange-PIE. M. Amir Mehmood

Curso de Telefonía IP para el MTC. Sesión 2 Requerimientos principales. Mg. Antonio Ocampo Zúñiga

Implementation of Video Voice over IP in Local Area Network Campus Environment

TECHNICAL CHALLENGES OF VoIP BYPASS

VoIP Testing IxChariot

Technote. SmartNode Quality of Service for VoIP on the Internet Access Link

Measurement of V2oIP over Wide Area Network between Countries Using Soft Phone and USB Phone

Comparison of Wireless Protocols. Paweł Ciepliński

A Generic Algorithm for Midcall Audio Codec Switching

VoIP QoS on low speed links

Network administrators must be aware that delay exists, and then design their network to bring end-to-end delay within acceptable limits.

MOS Technology Brief Mean Opinion Score Algorithms for Speech Quality Evaluation

Knowledge Is Power: Do what s best for the client.

Call Admission Control and Traffic Engineering of VoIP

Sync & Sense Enabled Adaptive Packetization VoIP

Padma Charan Das Dept. of E.T.C. Berhampur, Odisha, India

Authors Mário Serafim Nunes IST / INESC-ID Lisbon, Portugal mario.nunes@inesc-id.pt

Access the Test Here

CISCO IOS IP SERVICE LEVEL AGREEMENTS: ASSURE THE DELIVERY OF IP SERVICES AND APPLICATIONS

Encapsulating Voice in IP Packets

Department of MIIT, University of Kuala Lumpur (UniKL), Malaysia

Lecture 8 Performance Measurements and Metrics. Performance Metrics. Outline. Performance Metrics. Performance Metrics Performance Measurements

VOICE OVER IP AND NETWORK CONVERGENCE

Voice Over IP Performance Assurance

Performance Monitoring of VoIP with Multiple Codecs Using IPv4 and IPv6to4 Tunnelling Mechanism on Windows and Linux

A New Adaptive Redundancy Control Algorithm For VoIP Applications

A Quality of Experience based Approach for Wireless Mesh Networks*

An Introduction to VoIP Protocols

Voice over IP. Overview. What is VoIP and how it works. Reduction of voice quality. Quality of Service for VoIP

MINIMUM NETWORK REQUIREMENTS 1. REQUIREMENTS SUMMARY... 1

Objective Speech Quality Measures for Internet Telephony

Voice over Internet Protocol (VoIP) systems can be built up in numerous forms and these systems include mobile units, conferencing units and

VoIP. Wireless, P2P and New Enterprise Voice over IP

QOS Requirements and Service Level Agreements. LECTURE 4 Lecturer: Associate Professor A.S. Eremenko

SUNYIT. Reaction Paper 2. Measuring the performance of VoIP over Wireless LAN

VoIP Conformance Labs

Business case for VoIP Readiness Network Assessment

Service resiliency and reliability Quality of Experience Modelling requirements A PlanetLab proposal. PDCAT'08 - Dunedin December 1-4, 2008

White Paper. PESQ: An Introduction. Prepared by: Psytechnics Limited. 23 Museum Street Ipswich, Suffolk United Kingdom IP1 1HN

Voice, Video and Data Convergence > A best-practice approach for transitioning your network infrastructure. White Paper

Goal We want to know. Introduction. What is VoIP? Carrier Grade VoIP. What is Meant by Carrier-Grade? What is Meant by VoIP? Why VoIP?

VegaStream Information Note Considerations for a VoIP installation

VoIP over WiMAX: Quality of Experience Evaluation

Quality of Service (QoS) and Quality of Experience (QoE) VoiceCon Fall 2008

Simulative Investigation of QoS parameters for VoIP over WiMAX networks

ANALYSIS OF LONG DISTANCE 3-WAY CONFERENCE CALLING WITH VOIP

Assessing the quality of VoIP transmission affected by playout buffer scheme and encoding scheme

Internet Technology Voice over IP

SIP Trunking and Voice over IP

Agilent Technologies Performing Pre-VoIP Network Assessments. Application Note 1402

Priority Based Dynamic Rate Control for VoIP Traffic

Performance Evaluation of Next Generation Networks using OPNET Simulator

Requirements of Voice in an IP Internetwork

icall VoIP (User Agent) Configuration

VOIP under: WLAN g. VS Telephone Landline. ENSC 427 Team 1 Luke Dang tld@sfu.ca Jason Tsai kta2@sfu.ca Jeffrey Tam jta6@sfu.

Monitoring VoIP Call Quality Using Improved Simplified E-model

NCTU-VT: a freeware for wireless VoIP performance measurement

Network Simulation Traffic, Paths and Impairment

Voice-Over-IP. Daniel Zappala. CS 460 Computer Networking Brigham Young University

WhitePaper: XipLink Real-Time Optimizations

Active Monitoring of Voice over IP Services with Malden

Voice over IP: RTP/RTCP The transport layer

Measurement Challenges for VoIP Infrastructures

Simulation Based Analysis of VOIP over MANET

Online course syllabus. MAB: Voice over IP

IP SLAs Overview. Finding Feature Information. Information About IP SLAs. IP SLAs Technology Overview

Recommended QoS Configuration Settings for Rosewill RNX-AC750RT Wireless Router

Quality of Service Analysis of site to site for IPSec VPNs for realtime multimedia traffic.

New Models for Perceived Voice Quality Prediction and their Applications in Playout Buffer Optimization for VoIP Networks

Clearing the Way for VoIP

Performance Analysis Of Active Queue Management (AQM) In VOIP Using Different Voice Encoder Scheme

Network Performance Evaluation of Latest Windows Operating Systems

Adaptive Coding and Packet Rates for TCP-Friendly VoIP Flows

Transcription:

1. Title Performance Analysis Proposal Performance Analysis of P2P VoIP Applications 2. Objective To evaluate the dynamic behavior (i.e., CODEC usage) of VoIP P2P applications (e.g. Skype, Google Talk) under different network conditions and the impact on voice quality. 3. Research Questions Which s achieve better results? (this is not the same as which is the best? ) Which codec adaptations each perform on different network conditions (i.e., bandwidth, delay, jitter, packet loss)? How long does it take to adapt to a good codec? Which applications do cause more packet loss? Given that the network is low (e.g., 30-40K), would the application adapt to it, or the network quality metrics would reflect poor performance? (packet loss). 4. Environment Considerations Two users exchanging VoIP packets across a controlled network, which is a testbed network configured for different conditions using a WAN emulator (such as nist.net). The users must have access to the public Internet, in order to be able to be authenticated by the application, to have access to the contact list and to be able to start a voice call. However, the actual traffic goes through the emulated network. 5. Metrics Two different types of metrics, reflecting different aspects: End-user characteristics: MOS, E-model, PESQ, PSQM, PAMS, voice quality prediction? Network characteristics:, delay, jitter, packet loss. A different metric could be also used to evaluate TCP-friendliness of these s. In other words, at what extent does their adaptation (or the lack of it) harm (or not) TCP flows?

6. Experimental Design 6.1. Parameters Experiments assume two users, S (the sender) and R (the receiver) sending/receiving a fixed audio source over an emulated IP network. A fixed audio source must be used because depending on the sound source, Skype (and possibly others) generates different bit rates (this was observed in Net-Peeker, during a Skype session). 6.2. Metrics Metrics are the observable outcome of an experiment (also known as response variable). The following metrics are defined for these experiments, along with the method used for collecting them. a) The average size of packets delivered to R (measured on S) Packet sizes can be easily collected by reading the tcpdump file. b) The time the application takes to adapt after changing network conditions (assuming it redefines its strategy/choose another codec). A variant of this metric will be the initial adaptation delay, i.e., the time an application takes to find the correct codec/configuration for a given network condition. Adaptation delay is the time between two unambiguously different bitrate levels (or two average packet sizes, if we can get some insight from it, in case the application adapts by changing codec sampling or frame size). c) The PESQ MOS, as defined by the ITU recommendation P.862 ranging from 1.0 (worst) up to 4.5 (best). It represents the effect of network conditions on end-user audio perceived quality The PESQ MOS may be obtained by a method comprised of two phases: first, recording output sound in S and input sound in R; and second, submitting both sound files to the PESQ algorithm (C program freely available). d) Bitrate generated by the application (measured on S), used for comparing with the actual throughput observed on R and for computing the adaptation delay. Tcpstat computes the bitrate using the tcpdump file as input. e) Throughput, delay, jitter, packet loss (measured on R) Throughput: tcpstat Packet loss and jitter: ipstats Delay: ipstat + NTP

6.3. s and s s are variables that affect the outcome of the experiments. Each factor has a set of alternative values (levels). s are the values each factor can assume, that is, each level is an alternative for the correspondent factor. # 1 Skype, Google Talk, Yahoo Messenger with voice 2 Noisy (e.g. music), conversation (with periods of silence) 3 1M, 256K 150K, 56K, 28K, 9.6K 3 0, 1ms, 10ms, 100ms, 500ms, 1s, 10s 4 WAN packet loss 0.01%, 0.1%, 1%, 5%, 10%, 50% 5 WAN Jitter 0, 10%, 20% 6 TCP session (using TG?), 6.4. Experiments Next sections describe the specific experiments to be conducted. The factors and levels are described in the correspondent table and the observable values are all metrics defined on section 6.2. Since random variables are involved, in order to achieve good statistical accuracy it must be adopted a confidence level of 95% and a maximum error (precision) of 5%. Experiment duration: 1 hour Replications: 60 replications of 1 minute each. Two different types of experiments may be considered. In the first type the values of the levels are fixed, i.e., they are not varied during each experiment. In the second type, the levels are varied during the experiments. The adaptation delay can only be obtained from dynamic experiments. On the other hand, if only dynamic experiments are used, it may be difficult to collect the other metrics and associate them to a specific network condition. In other words, it may be difficult to automatically and unambiguously tell which metrics refers to each network conditions, because of the adaptation delay, that may vary depending on the type and value of network parameter that is changed (loss, delay,, etc). 6.4.1. Impact of bottleneck on selected metrics

10Mbps, 1Mbps, 100kbps, 56kbps, 28kbps, 9.6kbps Fixed = 100ms 6.4.2. Impact of delay on selected metrics Fixed = 128kbps 0, 1ms, 10ms, 100ms, 500ms, 1s, 10s 6.4.3. Impact of packet loss on selected metrics Fixed = 128kbps Fixed = 100ms WAN packet loss 0%, 1%, 5%, 10%, 50%

6.4.4. Impact of Jitter on selected metrics Fixed = 128kbps Fixed = 100ms WAN Jitter 0, 10%, 20% 6.4.5. TCP friendliness of the 10Mbps, 1Mbps, 100kbps, 56kbps, 28kbps, 9.6kbps Fixed = 1ms A (stable) TCP session 7. Research Group John Doe Jane Doe Joe Doe 8. Workplan Action 1 Environment Configuration

Activity 1.1 Installing Nist.Net in the NGN testbed Activity 1.2 Configuring routing and testing Action 2 Finding voice quality metrics Activity 2.1 Surveying voice quality metrics Activity 2.2 Comparing different metrics Activity 2.3 Choosing a particular metric Action 3 Background Survey Activity 3.1 Reading Skype paper Activity 3.2 Looking for more information on Skype, Google talk, etc Activity 3.3 Surveying the state of the art in similar papers (involving typical VoIP comparisons, for instance, SIP and H.323) Action 4 Performing experiments 9. Schedule Activity 4.1 Defining detailed experimental results Activity 4.2 Performing experiments Activity 4.3 Collecting and analyzing results Activity 4.4 Formatting and writing down results Activity 1.1 1.2 2.1 2.2 2.3 3.1 3.2 3.3 4.1 4.2 4.3 4.4 Week 1 2 3 4 5 6 7 8 9 10 11 12