1. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 1 Internet Telephony Shivkumar Kalyanaraman Based upon slides of Henning Schulzrinne (Columbia)

2. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 2 q Telephony: history and evolution q IP Telephony: Why ? q Adding interactive multimedia to the web q Being able to do basic telephony on IP with a variety of devices q Basic IP telephony model q Protocols: SIP, H.323, RTP, Coding schemes, MGCP, RTSP q Future: Invisible IP telephony and control of appliances Overview

3. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 3 Public Telephony (PSTN) History q 1876 invention of telephone q 1915 first transcontinental telephone (NY–SF) q 1920’s first automatic switches q 1956 TAT-1 transatlantic cable (35 lines) q 1962 digital transmission (T1) q 1965 1ESS analog switch q 1974 Internet packet voice q 1977 4ESS digital switch q 1980s Signaling System #7 (out-of-band) q 1990s Advanced Intelligent Network (AIN)

4. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 4 Telephone Service in the US AT&T Divestiture

5. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 5 Telephone System Overview q Analog narrowband circuits: home-> central office q 64 kb/s continuous transmission, with compression across oceans q  -law: 12-bit linear range -> 8-bit bytes q Everything clocked a multiple of 125 s q Clock synchronization  framing errors q AT&T: 136 “toll”switches in U.S. q Interconnected by T1, T3 lines & SONET rings  Call establishment “out-of-band” using packet- switched signaling system (SS7)

6. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 6 Telephony: Multiplexing  Telephone Trunks between central offices carry hundreds of conversations: Can’t run thick bundles!  Send many calls on the same wire: multiplexing  Analog multiplexing  bandlimit call to 3.4 KHz and frequency shift onto higher bandwidth trunk  Digital multiplexing: convert voice to samples  8000 samples/sec => call = 64 Kbps

7. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 7 Trends: Price of Phone Calls: NY - London AT&T Divestiture

8. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 8 Trends: Data vs Voice Traffic Since we are building future networks for data, can we slowly junk the voice infrastructure and move over to IP?

9. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 9 Trends: Phone vs Data Revenues

10. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 10 Private Branch Exchange (PBX) 7043 7040 7041 7042 External line Telephone switch Private Branch Exchange 212-8538080 Another switch Corporate/Campus Internet Corporate/Campus LAN Post-divestiture phenomenon...

11. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 11 External line 7043 7040 7041 7042 PBX Corporate/Campus Internet LAN 8154 8151 8152 8153 PBX Another campus LAN IP Telephony: PBX Replacement

12. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 12 Voice over Packet Market Forecast – North America

13. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 13 Invisible Internet Telephony q VoIP technology will appear in... q Internet appliances q home security cameras, web cams q 3G mobile terminals  fire alarms  chat/IM tools  interactive multiplayer games

14. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 14 IPtel for appliances: “Presence”

15. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 15 Taxonomy of Speech Coders Speech Coders Waveform CodersSource Coders Time Domain: PCM, ADPCM Frequency Domain: e.g. Sub-band coder, Adaptive transform coder Linear Predictive Coder Vocoder  Waveform coders: attempts to preserve the signal waveform not speech specific (I.e. general A-to-D conv) q PCM 64 kbps, ADPCM 32 kpbs, CVSDM 32 kbps qVocoders: q Analyse speech, extract and transmit model parameters q Use model parameters to synthesize speech q LPC-10: 2.4 kbps qHybrids: Combine best of both… Eg: CELP (used in GSM)

16. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 16 Speech Quality of Various Coders

17. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 17 Applications of Speech Coding q Telephony, PBX q Wireless/Cellular Telephony q Internet Telephony q Speech Storage (Automated call-centers) q High-Fidelity recordings/voice q Speech Analysis/Synthesis q Text-to-speech (machine generated speech)

18. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 18 Pulse Amplitude Modulation (PAM)

19. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 19 Pulse Code Modulation (PCM) * PCM = PAM + quantization

20. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 20 Quantization

21. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 21 Companded PCM Small quantization intervals to small samples and large intervals for large samples Excellent quality for BOTH voice and data Moderate data rate (64 kbps) Moderate cost: used in T1 lines etc

22. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 22 Companding

23. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 23 How it works for T1 Lines Companding blocks are shared by all 16 channels

24. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 24 Adaptive Gain Encoding Automatic Gain control (AGC), but accounting for silence periods

25. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 25 Time Waveform of Voiced/Unvoiced Sound High correlation (0.85) between samples, cycles, pitch intervals etc

26. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 26 Differential PCM Exploits sample-to-sample correlation (0.85) => differences require fewer bits; feedback avoids cascading quantization errors

27. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 27 Delta Modulation Used in first-generation PBXs (switching was more sensitive to Digital conversion cost and less sensitive to quality or data rate)

28. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 28 Adaptive Predictive Coding Adapt both the prediction coefficients (alphas) and the estimates Based upon past or present samples => 20 dB prediction gain

29. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 29 Subband Coding Frequency domain analysis of input instead of time-domain Analysis: adjust quantization based upon energy level of each band Eg: G.722 coder: 7kHz voice w/ 64 kbps

30. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 30 G.722 (7 kHz) audio Codec

31. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 31 Recall: Taxonomy of Speech Coders Speech Coders Waveform CodersSource Coders Time Domain: PCM, ADPCM Frequency Domain: e.g. Sub-band coder, Adaptive transform coder Linear Predictive Coder Vocoder  Waveform coders: attempts to preserve the signal waveform not speech specific. q PCM 64 kbps, ADPCM 32 kpbs, CVSDM 32 kbps qVocoders: q Analyse speech, extract and transmit model parameters q Use model parameters to synthesize speech q LPC-10: 2.4 kbps qHybrids: Combine best of both… Eg: CELP

32. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 32 Vocoders Encode only perceptually important aspects of speech w/ fewer bits than waveform coders: eg: power spectrum vs time-domain accuracy

33. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 33 LPC Analysis/Synthesis

34. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 34 Speech Generation in LPC

35. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 35 Multi-pulse LPC

36. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 36 CELP Encoder

37. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 37 Example: GSM Digital Speech Coding q PCM: 64kbps too wasteful for wireless q Regular Pulse Excited -- Linear Predictive Coder (RPE--LPC) with a Long Term Predictor loop. q Subjective speech quality and complexity (related to cost, processing delay, and power) q Information from previous samples used to predict the current sample: linear function. q The coefficients, plus an encoded form of the residual (predicted - actual sample), represent the signal. q 20 millisecond samples: each encoded as 260 bits =>13 kbps (Full-Rate coding).

38. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 38 Speech Quality of Various Coders

39. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 39 Speech Quality (Contd)

40. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 40 VoIP Camps ISDN LAN conferencing IP H.323 I-multimedia WWW IP SIP Call Agent SIP & H.323 IP “Softswitch” BISDN, AIN H.xxx, SIP “any packet” BICC Conferencing Industry Netheads “IP over Everything” Circuit switch engineers “We over IP” “Convergence” ITU standards Our focus

41. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 41 Internet Multimedia Protocol Stack

42. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 42 IP Telephony Protocols: SIP, RTP  Session Initiation Protocol - SIP  Contact “office.com” asking for “bob”  Locate Bob’s current phone and ring  Bob picks up the ringing phone  Real time Transport Protocol - RTP  Send and receive audio packets

43. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 43 Internet Telephony Protocols: H.323

44. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 44 H.323 (contd) q Terminals, Gateways, Gatekeepers, and Multipoint Control Units (MCUs)

45. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 45 H.323 vs SIP IP and lower layers TCP UDP TPKT Q.931 H.245 RASRTCP RTP Codecs Terminal Control/Devices Transport Layer SIPSDP RTP Codecs RTCP Terminal Control/Devices Typical UserAgent Protocol stack for Internet

46. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 46 SIP vs H.323 q Text based request response q SDP (media types and media transport address) q Server roles: registrar, proxy, redirect q Binary ASN.1 PER encoding q Sub-protocols: H.245, H.225 (Q.931, RAS, RTP/RTCP), H.450.x... q H.323 Gatekeeper - Both use RTP/RTCP over UDP/IP - H.323 perceived as “heavyweight”

47. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 47 Light-weight signaling: Session Initiation Protocol (SIP) q IETF MMUSIC working group q Light-weight generic signaling protocol q Part of IETF conference control architecture: q SAP for “Internet TV Guide” announcements q RTSP for media-on-demand q SDP for describing media q others: malloc, multicast, conference bus,... q Post-dial delay: 1.5 round-trip time (with UDP) q Network-protocol independent: UDP or TCP (or AAL5 or X.25)

48. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 48 SDP: Session Description Protocol q Not really a protocol – describes data carried by other protocols q Used by SAP, SIP, RTSP, H.332, PINT. Eg: v=0 o=g.bell 877283459 877283519 IN IP4 132.151.1.19 s=Come here, Watson! u=http://www.ietf.org e=g.bell@bell-telephone.com c=IN IP4 132.151.1.19 b=CT:64 t=3086272736 0 k=clear:manhole cover m=audio 3456 RTP/AVP 96 a=rtpmap:96 VDVI/8000/1 m=video 3458 RTP/AVP 31 m=application 32416 udp wb

49. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 49 SIP functionality q IETF-standardized peer-to-peer signaling protocol (RFC 2543):  Locate user given email-style address q Setup session (call) q (Re)-negotiate call parameters q Manual and automatic forwarding q Personal mobility: different terminal, same identifier q Call center: reach first (load distribution) or reach all (department conference) q Terminate and transfer calls

50. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 50 SIP Addresses Food Chain

51. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 51 SIP components q UAC: user-agent client (caller application) q UAS: user-agent server à accept, redirect, refuse call q redirect server: redirect requests q proxy server: server + client q registrar: track user locations  user agent = UAC + UAS  often combine registrar + (proxy or redirect server)

52. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 52 Are true Internet hosts Choice of application Choice of server IP appliances Implementations 3Com (3) Columbia University MIC WorldCom (1) Mediatrix (1) Nortel (4) Siemens (5) 4 IP SIP Phones and Adaptors 1 3 Analog phone adaptor Palm control 2 54

53. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 53 SIP-based Architecture SIP proxy, redirect server SQL database sipd e*phone sipc Software SIP user agents Hardware Internet (SIP) phones SIPH.32 3 converto r NetMeeting sip323 H.323 rtspd SIP/RTSP Unified messaging RTSP media server sipum Quicktime RTSP clients RTSP SIP conference server sipconf T1/E1 RTP/SIP Telephone Cisco 2600 gateway Telephone switch Web based configuration Web server

54. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 54 SIP proxy, redirect server SQL database sipd e*phone sipc Software SIP user agents Hardware Internet (SIP) phones Web based configuration Web server Call Bob Example Call Bob signs up for the service from the web as “bob@ecse.rpi.edu” He registers from multiple phones Alice tries to reach Bob INVITE ip:Bob.Wilson@ ecse.rpi.edu sipd canonicalizes the destination to sip:bob@ecse.rpi.edu sipd rings both e*phone and sipc Bob accepts the call from sipc and starts talking ecse.rpi.edu

55. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 55 PSTN to IP Call PBX PSTN External T1/CAS Regular phone (internal) Call 9397134 1 SIP server sipd Ethernet 3 SQL database 4 7134 => bob sipc 5 Bob’s phone Gateway Internal T1/CAS (Ext:7130-7139) Call 7134 2

56. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 56 IP to PSTN Call Gateway (10.0.2.3) 3 SQL database 2 Use sip:85551212@10.0.2.3 Ethernet SIP server sipd sipc 1 Bob calls 5551212 PSTN External T1/CAS Call 5551212 5 5551212 PBX Internal T1/CAS Call 85551212 4 Regular phone (internal, 7054)

57. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 57 Traditional voice mail system Alice 939-7063 Bob 853-8119 Dial 853-8119 Phone is ringing.. The person is not available now please leave a message...... Your voice message... Disconnect Bob can listen to his voice mails by dialing some number.

58. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 58 SIP-based Voicemail Architecture INVITE bob@office.com Alice phone1.office.com Bob Alice calls bob@office.com through SIP proxy. SIP proxy forks the request to Bob’s phone as well as to a voicemail server. vm.office.com The voice mail server registers with the SIP proxy, sipd INVITE bob@vm.office.com INVITE bob@phone1.office.com REGISTER bob@vm.office.com

59. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 59 Voicemail Architecture v-mail rtspd Alice vm.office.com; 200 OK CANCEL SETUP RTP/RTCP phone1.office.com; Bob After 10 seconds vm contacts the RTSP server for recording. vm accepts the call. Sipd cancels the other branch and......accepts the call from Alice. Now user message gets recorded

60. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 60 SIP-H.323: Interworking Problems Eg: Call setup translation Q.931 SETUP Q.931 CONNECT INVITE 200 OK ACK Terminal Capabilities Open Logical Channel H.323SIP Destination address (Bob@office.com) Media capabilities (audio/video) Media transport address (RTP/RTCP receive) H.323: Multi-stage dialing

61. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 61 MGCP and Megaco q Media Gateway Controller Protocol (RFC 2705) q Controlling Telephony Gateways from external call control elements called media gateway controllers (MGC) or call agents q Gateways: Eg: RGW : physical interfaces between VoIP network and residences q Call control "intelligence" is outside the gateways and handled by external call control elements q Goal: scalable gateways between IP telephony and PSTN q Successor to MGCP: H.248/Megaco

62. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 62 MGCP Architecture RGW: Residential Gateway TGW: Trunk Gateway Goal: large-scale phone-to-phone VoIP deployments

63. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 63 Summary q Telephony and IP Telephony q Protocols: SIP, SDP, H.323, MCGP q Example operation and services: q Calls, voice mail etc q Future: Integration with Web and long-term replacement for current telephone systems