Master’s Thesis: Performance of Multimedia Services in Packet Switched Mobile Networks Sami Ehoniemi sami.ehoniemi@teliasonera.com 24 July, 2016 1 Basic Information • Title: Performance of Multimedia Services in Packet Switched Mobile Networks • Supervisor: Professor Timo Korhonen • Instructor: Mika Raitola, M.Sc. • Place of work: TeliaSonera Finland – Emerging Technologies and Innovations 24 July, 2016 2 Presentation Outline • Motivation & Background • Objectives & Methodology • Thesis Contents • Measurements • Conclusions 24 July, 2016 3 Motivation & Background • GPRS brought Packet Switched data services to GSM networks – Several new services possible • 3G UMTS developing towards an all IP network – IMS platform • Past ARPU from voice calls and SMS • Future growth from value added services incl. multimedia and messaging services • Performance acceptance of multimedia services a major challenge • Operators concerned with the proper functioning of the services and the effective use of the network 24 July, 2016 4 Objectives & Methodology • Objective was to evaluate: – GPRS network possibilities, challenges and problems – Functioning and performance of a couple of services: MMS, Photoshare, IMS PoC, and Fastchat – Effects of the services to the mobile network (GPRS, EDGE) • Methodology: – literary study based on 3GPP, OMA, and MENSA specifications, several books and conference articles – Measurements of several services 24 July, 2016 5 Thesis Contents 1. Introduction 2. Introduction to Mobile Networks 3. Overview of Messaging Services 4. Factors Affecting MMS Performance 5. IP Multimedia Subsystem (IMS) 6. IMS Case Study: MENSA PoC 7. Measurements 8. End-user Performance Analysis 9. Conclusions 24 July, 2016 6 Measurements (1/2) • MMS performance tests with several GPRS and EDGE terminals – MMS transmission divided into several time periods (GPRS connection delays, UL and DL transmission delay, MMSC delay) – UL and DL transmission rates – MMS roaming tests • IMS Photoshare performance tests with GPRS terminals – Comparison to MMS – Session establishment and image transmission delays – Statistics of HTTP and SIP traffic 24 July, 2016 7 Measurements (2/2) • IMS Push-to-talk over Cellular (PoC) – Two test sessions with GPRS and EDGE terminals – Session 1: Understanding of the overall PoC traffic – Session 2: MENSA consortium specified delay values (Session control establishment, floor control, voice transfer) • Fastchat Push-to-talk service – Comparison to MENSA consortium specified PoC – Voice delay values measured 24 July, 2016 8 Conclusions (1/3) • GPRS presents a challenging environment for achieving good application performance – Long and variable network delay – Low and variable network throughput – Difficult for interactive services • EDGE brings capacity and throughput improvements, but not delay improvements • UMTS provides much smaller network latency and improved throughput 24 July, 2016 9 Conclusions (2/3) • Main factors affecting the MMS performance: – Terminal Multislot Class (maximum number of simultaneous GPRS or EDGE timeslots), dedicated GPRS capacity – Additional transmission throughput with EDGE – Transport protocol stack (WAP 2.0 better than WAP 1.2.1) – Always GPRS attached mode cut the GPRS connection time – Notification method (SMS notification slow) • Longer GPRS connection and UL transmission delays with a roaming terminal • Measured delays within the proper limits for a non-real-time service 24 July, 2016 10 Conclusions (3/3) • Photoshare an effective way to share media (images, etc.) • Image transmission from terminal to terminal faster than with MMS – High GPRS connection delays of MMS – Long notification delay of MMS • MENSA specified PoC delays achievable (low variance) • SIP signaling compression highly justified • IP/UDP (RFC2507) and RTP/UDP/IP (ROCH RFC3095) header compression methods should be examined • EDGE brings capacity improvement and session control • Need for GPRS QoS with PoC • Fastchat voice delays longer due to long server delay 24 July, 2016 11 Thank You! 24 July, 2016 12