More Related Content
Similar to The VoLTE User Experience--Better or Worse (20)
The VoLTE User Experience--Better or Worse
- 1. The VoLTE User Experience Better or Worse?
- 2. 2 Challenge VoLTE is Different Quantify The VoLTE User Experience Solve Making VoLTE Better (not Worse) Outline
- 3. 3 Sources Facts Q1 2013: 6.4 billion mobile subscribers, with smartphones comprising 50% of new mobile device sales Voice and SMS represent approximately 70% of all global wireless revenues Voice-over-LTE (VoLTE) revenues estimated to reach $2 billion by 2016 Trends Highly competitive market, ARPU no longer increasing HD Voice is one way to seek competitive advantage (73 networks so far) Voice and other services moving to more efficient all-IP LTE networks • Ericsson Mobility Report – On the Pulse of the Networked Society – June 2013 • Global Mobile Suppliers Association – Mobile HD Voice: Global Update Report – June 4th, 2013
- 4. 4 • How can I compare the VoLTE voice experience offered by different suppliers and technology options? • How can I know that new infrastructure will deliver high-quality VoLTE services, before it goes live? • How can I ensure that my devices offer the VoLTE voice experience my customers expect? • How can I ensure that my network offers the VoLTE voice experience my customer expect? • How does the VoLTE Voice experience on my network and devices compare to my competitors? VoLTE: Better or Worse?
- 5. 5 VOLTE IS DIFFERENT The VoLTE User Experience: Better or Worse?
- 6. 6 Physical Data Link Network Transport Session Presentation Application Physical Data Link Network Transport Session Presentation Application SIP, RTP, RTCP TCP, UDP, SCTP IP, RRC, NAS Voice Service Applications MAC, RLC, PDCP OFDM, WCDMA, etc. Mobile Device Network2G/3G Dedicated Voice Bearers – Layer 1-3 Legacy Voice Services Depend on Robustness of Layers 1-3
- 7. 7 Physical Data Link Network Transport Session Presentation Application Physical Data Link Network Transport Session Presentation Application SIP, RTP, RTCP TCP, UDP, SCTP IP, RRC, NAS Voice Service Applications MAC, RLC, PDCP OFDM, WCDMA, etc. Mobile Device Network4G Dedicated Data Bearers – Layer 1-7 4G Voice Services Depend on Robustness at All Layers
- 8. 8 QoS RAN IMS Codec Dedicated vs. Non-Dedicated Bearers Semi-persistent scheduling resource allocation Session Initiation Protocol (SIP) HD Voice (Wideband AMR) Quality of Service Class Identifier (QCI) Transmission Time Interval (TTI) Bundling Policy and Charging Rules Function (PCRF) Dynamic scheduler in eNodeB Robust Header Compression (RoHC) Real-time Transport Protocol (RTP) Real-time Streaming Protocol (RTSP) Key Enablers of VoLTE Fall Into Four Categories
- 9. 9 QoS Enables VoLTE Packets to be Prioritized
- 10. 10 QCI Sets Specific Packet Loss & Delay Targets Quality of Service (QoS) Class Identifier indicates max delay & packet error rates for each LTE bearer Various network functions rely on the QCI to make prioritization & resource allocation decisions
- 11. 11 eNB scheduler dynamically allocates resource blocks (RBs) across all users. QoS delay budgets are a key factor in allocations eNB Dynamic Scheduler Grants Resources
- 12. 12 Semi-Persistent Resource Allocation for VoLTE Semi-persistent scheduling reduces the signaling overhead for RB allocation by granting periodic use of a set of RBs VoLTE sends short packets on a regular basis: SPS ensures resources are available which match the periodicity of VoLTE
- 13. 13 16QAM Signal Constellation (4 bits per symbol) 64QAM Signal Constellation (6 bits per symbol) Link adaptation manages the packet loss rate by adjusting the rate at which bits are transmitted up or down based on radio conditions (SNR). MoreBits/s&HigherSNR Link Adaptation is Key to Packet Loss
- 14. 14 HARQ & TTI Bundling Impact Packet Loss & Delay HARQ Reduces Packet Loss / Increases Delay (Jitter) TTI Bundling Reduces HARQ Latency
- 15. 15 IP Multimedia Subsystem Session Initiation Protocol (SIP) Session Bit rate Packet size Packet transport frequency RTP payload Bandwidth adaptation Policy and Charging Rules Function (PCRF) Policy rules (bandwidth, quality class, IP packet filters) PDN Gateway (P-GW) Interpret rules and establish EPC dedicated bearers for voice
- 16. 16 HD Voice (WB-AMR Codec) Type of Channel Channel Bandwidth Type of Service Narrowband 50Hz – 3.8KHz Narrowband Voice Wideband 50Hz – 7.5KHz HD Voice (including VoLTE) Super-wideband 20Hz – 14KHz HD Voice (including VoLTE) Channel bandwidths for different voice services
- 17. 17 THE VOLTE USER EXPERIENCE The VoLTE User Experience: Better or Worse?
- 18. 18 End User Experience Key Performance Indicators (KPIs) Ability to make and maintain calls Call Initiation Rate (%), Call Drop Rate (%) Time it takes for a phone to start ringing Call Setup Time (s) Speech quality during a call Mean Opinion Score (MOS) Mouth-to-ear delay or latency variations during a call Mouth-to-ear delay (s), Latency, jitter What to Measure?
- 19. 19 Packet Loss & Delay Drive QoE for VoLTE VoLTE packet loss directly impacts speech quality Source: “Validating voice over LTE end-to-end”, Ericsson Review, January 2012. Speech Quality vs. Frame Error Rate Packet loss rate of < 1% results in good speech quality on avg.
- 20. 20 Packet Loss & Delay Drive QoE for VoLTE VoLTE packet delay impacts mouth- to-ear delay, and packet delay variability (jitter) impacts both speech quality & mouth-to-ear delay Speech Quality vs. Mouth to Ear Delay Source: “Validating voice over LTE end-to-end”, Ericsson Review, January 2012. ITU recommended end- to-end mouth-to-ear delay is <200 ms
- 21. 21 UE-Specific Delays Dominate Mouth-to-Ear Delay Packet delay is strongly impacted by LTE scheduling & HARQ: • Average E2E transport delay • Jitter buffer delay Breakdown of mouth-to-ear delay for VoLTE lab and field tests performed by Ericsson Source: “Validating voice over LTE end-to-end”, Ericsson Review, January 2012.
- 22. 22 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% PercentageofValuesinMOSRange MOS Range MOS Distribution Both Speech Paths NB/Standard Mode AMR-WB enabled Source: Spirent Testing on Live Network with Nomad HD Speech Quality Impacted by Codec
- 23. 23 PESQ (ITU-T P.862) POLQA (ITU-T P.863) Codecs AMR EFR AMR AMR-WB EFR EVRC EVRC-B EVRC-WB Reference Speech Material (sampling frequency) 8 kHz 8 kHz 48 kHz Applications POTS VoIP 3G HD Voice Voice Enhancement Devices POLQA has more robust quality predictions for … Cross-technology quality benchmarking (e.g., GSM vs. CDMA) Noise reduction and voice quality enhancement Time-scaling, unified communication and VoIP Filtering and spectral shaping Recordings made at an ear simulator Perceptual Evaluation of Speech Quality (PESQ) vs. Perceptual Objective Listening Quality Analysis (POLQA)
- 24. 24 Downlink MOS Uplink MOS Device Device A Device B Device C Device D Device A Device B Device C Device D Average 3.09 3.34 3.16 3.62 3.46 3.81 3.44 3.31 Standard Deviation 0.42 0.08 0.39 0.29 0.22 0.03 0.27 0.12 Maximum Score 3.56 3.45 3.43 3.84 3.65 3.85 3.83 3.45 % MOS less than 3.0 33% 0% 17% 0% 0% 0% 0% 0% POLQA MOS Score Comparison for VoLTE Devices Source: Spirent Testing on Live Network with Nomad HD
- 25. 25 Call Initiation and Setup Time Comparison for VoLTE Devices Source: Spirent Testing on Live Network with Nomad HD
- 26. 26 Call Drop Comparison for VoLTE Devices Source: Spirent Testing on Live Network with Nomad HD
- 27. 27 MAKING VOLTE BETTER The VoLTE User Experience: Better or Worse?
- 28. 28 Quantify, Ensure, Measure, Verify, Troubleshoot, Benchmark … Test
- 29. 29 Measuring device performance across multiple OS’s and technologies Testing in both live and simulated network environments Evaluating multiple devices simultaneously Managing extensive testing projects from a centralized location Measuring how a device’s performance will impact subscribers Comparing VoLTE voice quality to circuit switched voice: Is it as good or better? Common VoLTE Test Challenges
- 30. 30 Metrics that focus on what the end-users experience (including speech quality, the ability to make and maintain calls, and mouth-to-ear latency) Use the same voice service measurement systems in the field and in the lab Measurement systems that can test any device, on any network, anywhere in the world and still provide one central location for results collection and analysis Lab test solutions that provide simple interfaces for LTE and VoLTE configuration while also enabling fast creation of automated VoLTE tests Coverage of relevant compliance tests originating from operators and standards organizations Characteristics of Ideal Test Strategy
- 31. 31 $$ $$$$ Test Early and Often, in Lab & Live Lab (Emulated Network) Live
- 32. 32 Spirent Believes in “Better” How can we help you improve VoLTE quality and time-to-market?
- 33. 33 For More Information Spirent’s VoLTE and HD Voice web page: • Application Notes for CS8 and Nomad HD Voice Quality testing • Video demonstrations http://www.spirent.com/go/VoLTE
- 34. 34
Editor's Notes
- #20: LTE air-interface is a key contributor to end-to-end packet loss and delayThe level of jitter requires an increased jitter buffer size – that increases the delay and may lead to dropped packets for excessively delayed packets
- #21: LTE air-interface is a key contributor to end-to-end packet loss and delayThe level of jitter requires an increased jitter buffer size – that increases the delay and may lead to dropped packets for excessively delayed packets