Change search
ReferencesLink to record
Permanent link

Direct link
ARAM WebRTC: A Rate Adaptation Model for WebRTC Real-Time Interactive Video Over 3GPP LTE
2013 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Nowadays, interactive audio/video communication is available even on mobile phone browsers using the WebRTC API (Real Time Communication on Web) and it is the state of the art of real time communication. WebRTC is developed as part of the RTCWeb protocol, which standardizes the real time communication across browsers in order to avoid failures in session set ups and enhance the session’s Quality of Experience (QoE).Just as previous real-time technologies, WebRTC requires a certain QoE from the networks, such as a certain available bandwidth capacity, packet delay and jitter. However, the Internet infrastructure still relies on the shared link and exorbitant numbers of sources might send data at high rates, which can lead to a congestion bottleneck problem during a real time session resulting in a lower service quality and end-user QoE. Furthermore, the congestion might occur at any time in a wireless network, due to varying available bandwidth capacity in relation with a multitude of mobile users.Therefore, in order to deal with the bottleneck problem and optimize performance of real time video of WebRTC, this thesis presents, ARAM WebRTC, a rate adaptation model that provides a fair available bandwidth allocation for WebRTC applications. It is based on real time available bandwidth estimation and measurements using the latest generation of wireless access networks, LTE (3GPP Long Term Evolution).The ARAM WebRTC model exploits two different available bandwidth capacity estimation algorithms in tandem. The first one is the Kalman filtering algorithm that is used to compute states of available bandwidth capacity, ”Under-Using” and “Over-Using”, based on video frame delays as described in the WebRTC Google Congestion Control Algorithm draft. The other is the TFRC (TCP Friendly rate control) algorithm which calculates available bandwidth capacity based on packet loss rates. TFRC is employed as a post available bandwidth capacity estimation technique to provide fairness between TCP and the real time interactive video traffic flows.Measurements and evaluations have been carried out using a Java based system simulator and the Matlab tool. In order to observe the performance of the presented rate adaptation model, the number of active users was varied in three different simulation scenarios, namely single Video traffic, mixed Video-VoIP traffic, and mixed Video-VoIPWeb traffic. The traffic scenarios have been simulated where Active Queue Management (AQM) was turned on and off, while the data traffic was transmitted over single and multi LTE QoS bearers. Also, uplink and downlink scenarios were simulated separately in order to decrease the number of dependences.The results show that the ARAM WebRTC model provides a fair bandwidth allocation between the two different traffic flow types: TCP and real time media traffic flows. Moreover, the simulation results also underline that excessive video frame delays and packet losses are prevented with the presented rate adaptation model of WebRTC interactive video. Hence, consequently the model improves performance and perceived quality of WebRTC real time interactive video.

Place, publisher, year, edition, pages
Keyword [en]
Technology, WebRTC, VoIP, video, UDP, TCP, RTCP, RTP, TFRC, 3GPP, LTE, realtime, adaptation, QoS, communication, Kalman, rate, congestion control
Keyword [sv]
URN: urn:nbn:se:ltu:diva-58457Local ID: f0a942e3-450f-4288-a91c-c5146b9bcaa2OAI: diva2:1031845
External cooperation
Subject / course
Student thesis, at least 30 credits
Educational program
Computer Science and Engineering, master's level
Validerat; 20131204 (global_studentproject_submitter)Available from: 2016-10-04 Created: 2016-10-04Bibliographically approved

Open Access in DiVA

fulltext(3185 kB)0 downloads
File information
File name FULLTEXT02.pdfFile size 3185 kBChecksum SHA-512
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Kilinc, Caner

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

ReferencesLink to record
Permanent link

Direct link