DESIGN AND IMPLEMENTATION OF VIDEO STREAMING APPLICATION FOR SLOW-

SPEED LINK - REALSTREAMER

Dr. Ayodeji O. Oluwatope

Lead, Network Utilisation Maximisation Sub-Group

Comnet Lab, Department of Computer Science and Engineering

Obafemi Awolowo University, Ile-Ife

url: comnet.oauife.edu.ng

aoluwato@oauife.edu.ng

Presentation Agenda

• Introduction - Project Goal, Project Benefits

• Research Methodology• OVSM – Rate Adaptive Scheme• OVSM Analysis• Simulation Results & Discussions• Implementation• Conclusion• References

Introduction

• Information theory (Shannon,1948)• Rate, R < Capacity, C - desireable• Rate, R ≥ Capacity, C - undesireable

• What is a slow-speed link?• Uplink (< 1Mbps) / download rate ( <4Mbps)

• What is a rate adaptive system?• high speed and slow speed difference compensator

Introduction contd

• Unbearable per-unit cost of Internet bandwidth

• Low level of internet users utilisation

• Classroom over population

Project Goals• create pluggable software(s) capable of streaming

multimedia traffics across slow-speed networks

• contribute to local contents development in the area of appropriate software engineering

• provide research opportunities for Masters and Doctorates Theses,

• produce high quality publications in reputable journals, and produce products patents

Project Benefits• creation of an appropriate software technology to reduce Internet

speed mis-match effect

• eradication of the digital divide currently subsisting between the upper and the lower class of the society

• increase deployment rate of value-added services via Internet such as telemedicine, e-learning, video surveillance etc.

• contribute to the strategic goal of Obafemi Awolowo University (OAU), Ile-Ife Science and Technology Park in the spinning-off SMEs

• improve on Nigerian Universities’ relevance to addressing our socio-economic challenges

Research Methodology• Approaches

• Analysis, empirical & prototyping

• Problem analysis phase • critical factors responsible for large end-to-end delay in slow-speed networks

were investigated,• existing video streaming systems auditing, stakeholders needs analysis,

industry standard and regulations gathering and harmonisation and requirement validation.

Research Methodoloy contd• Modelling phase

• End-to-end video frame scheduling scheme was developed and analysed for performance using Matlab

• Further performance analysis was carried out at packet-level using a discrete-event simulator (NS3).• average packet delay, packet-loss rate, bit error rate, BER, throughput, & jitter

in a hypothetical slow-speed network.

• Prototyping phase• Application design and prototyping using object-oriented design

approach and Java technology respectively.

OVSM- New Rate Adaptive Scheme

(Olayiwola and oluwatope ,2014)

OVSM Stead-State Analysis• Queue Q1 modelling

• M/D/1/K (olayiwaola et al., 2013)

• Queue Q2 modelling• D/D/1/K (olayiwaola et al., 2013)

Simulations Results from Matlab

olayiwaola et al., 2013)

Simulations Results from Matlab -contd

olayiwaola et al., 2013)

Simulations Results from ns3

Olayiwola and oluwatope, 2014

Realstreamer - Encoder

Realstreamer –Encoder

• Images Capturing is achieved using Sarxos WebCap Capture (http://webcam-capture.sarxos.pl)

• Java Audio was use in capturing audio data

• The opening and reading of media file are achieved using Xuggle API(www.xuggle.com).

Realstreamer –Encoder contd

• The Encoder is design in such a way to handle both prerecorder media and real time captured data.

• Both Audio and Video producers are interfaced to a buffer to handle race condition and thread related issues.

• Connections to the server are managed by EncoderStream Agent implemented with nettyio

Realstreamer – Server

Realstreamer –Decoder

Realstreamer –Decoder

• Decoder – The decoder gets data from the server transmits log data to the server, update image decoder and audio decoder.

• Image decoder gets the image, decodes it and updates the UI.

• Audio decoder gets the audio and play out.

Sample UI Designs

Currents Issues

• Performance bugs – high packet delay and packet loss rate

• Software usability, user-friendliness etc

Conclusion• Optimised video streaming model (OVSM) for multimedia

traffic formulated, modelled, and analysed

• OVSM is being adjusted for implementation

• Realstreamer architecturing has been achieved

• Realstreamer implementation is being perfected – OSS

• OVSM software real-life testing has commenced

Team Members• Researchers

• Dr. A. O. Oluwatope - Computer Science, OAU• Professor E. R. Adagunodo -Computer Science, OAU• Professor G. A. Adarounmu - Computer Science, OAU• Dr. Yinka Adesina – Mathematics, OAU• Dr. Ishaya Tanko - Computer Science, UH

• Research Assistants• Miss O. E. Olayiwola – Doctoral student• Mr D. Oyewo - Doctoral student

• Technical Assistant• Mr Azeez Isau - Bachealor Student

• Grants• TetFund Intervention Project (2013-2014)• OAU counterpart funding

References

• A.O. Oluwatope, D. T. Oyewo, F. E. Olayiwola, G. A. Aderounmu, and E.R. Adagunodo (2013) Buffer Occupancy of Double-Buffer Traffic Shaper in Real-Time Multimedia Applications Across Slow-Speed Links. To Appear in Communications and Network, Scientific Research Publishing, Inc. USA Vol 5 Issue 1

• Oluwafolake Olayiwola, Damilola Oyewo, Ayodeji Oluwatope, Adesola Aderounmu and Rotimi Adagundo. (2012). Double-Buffer Traffic Shaper Modelling For Multimedia Applications in Slow Speed Network. Journal of Issues in Informing Science and Information Technology, USA 9: 361-368.

• Oluwafolake Olayiwola and Ayodeji Oluwatope (2014).Packet-level Simulation of Real-Time Video in Slow-Speed Environment using NS-3. To Appear in IASTED Proceeding of Modelling and Simulation. 1-3 September 2014, Gaborone, Botswana

•Thank you for listening