An Introduction to Voice over IP

1
An Introduction to Voice over IP

• Lance Parr
• Lead Systems Administrator
• VoIP Laboratory Manager
• Texas AM University
• College Station, Texas

2
What is VoIP?

• VoIP (Voice over Internet Protocol), sometimes
  referred to as Internet telephony, is a method of
  digitizing voice, encapsulating the digitized
  voice into packets and transmitting those packets
  over a packet switched IP network.

3
Voice over IP - the basics

• Most implementations use H.323 protocol
• Same protocol that is used for IP video.
• Uses TCP for call setup
• Traffic is actually carried on RTP (Real Time
  Protocol) which runs on top of UDP.

4
VoIP Protocols

• H.323 Multimedia Standard
• H.225 RAS - Registration, Admission, Status
• Q.931 - Call Signaling (Setup Termination)
• H.245 - Call Control (Preferences, Flow Control,
  etc.)
• Lots of G.7XX CODECS for audio
• SIP Session Initialization Protocol
• Covered in next presentation

5
Heres how it stacks up
6
How they fit in The ISO Model
7
RTCP
RTP
UDP
TCP
Layer 3 IP
Layer 2 Ethernet/ATM
8
Comparison of Packet vs. Circuit Switching
9
H.323

• Definition a multimedia standard that provides a
  foundation to transport voice, video and data
  communications in an IP based non-QOS network.
• H.323 Zone
• Collection of terminals, gateways, MCUs
  registered with a single gatekeeper.

10
H.323 Entities

• Terminals (LAN Endpoints)
• Gateways (Optional but really useful)
• Gatekeepers (Also optional)
• MCUs

11
H.323 Equipment

• Gateway
• Device that connects H.323 voice network to
  non-H.323 voice network (SIP or PSTN)
• Allows H.323 terminals to communication with
  non-H.323 terminals
• Gatekeeper
• Provides address translation (H.323 E.164 to
  IP)
• Admission control for H.323 terminals and
  gateways
• Manage bandwidth allocation
• Other optional services

12
H.323 Equipment

• MCU (multipoint control unit)
• MC multipoint controller
• Routes call and control signaling to ensure
  endpoint compatibility
• MP multipoint processor
• Switches, mixes and processes vice and video
  streams to conferencing equipment

13
H.323 Equipment

• Terminal
• An endpoint that supports 2-way streaming with
  another H.323 terminal or gateway
• Originates and terminates calls
• Includes videoconferencing stations, hard phones,
  soft phones

14
Call Setup using H.225 RAS

• Registration, Admission and Status (RAS), is
  responsible for registration, admission, and
  disengaging procedures between H.323 Gatekeeper
  and Gateway.
• Discovery GRQ, GCF, GCR
• Unicast Discovery using UDP port 1718. Endpoint
  knows GK IP register directly
• Multicast using UPD multicast address 224.0.1.41
  non static, less admin overhead

15
Call Setup using H.225 RAS

• Registration by terminals, Gateways MCUs using
  H.323 ID or E.164 address
• RRQ Registration Request
• RCF Registration Confirm
• RRJ Registration Reject
• URQ Un-registration Request
• URF Un-registration Confirm
• URJ Un-registration Confirm

16
H.323 H.225 RAS Messages

• LRQ location request
• Gatekeeper A requests contact information from
  directory gatekeeper.
• LCF location confirm
• Gatekeeper B returns IP address of destination
  gateway to gatekeeper A.

17
Signaling using Q.931 messages

• Q.931 is a signaling protocol used to setup,
  manage, and terminate H.323 connections between
  endpoints.
• ARQ, ACF, ARJ Admission messages
• LRQ, LCF, LRJ Location Request messages
• IRQ, IRR, IACK, INAK Status messages
• BRQ, BCF, BRJ, RAI, RAC Bandwdith messages

18
H.323 H.225 RAS Messages

• ARQ admission request
• Gateway A requests admission to make a call.
• ACF admission confirm
• Gatekeeper A responds with IP address of
  destination gateway.

19
H.323 H.225 RAS Messages

• Request in Progress
• RIP
• Bandwidth change
• BRQ, BCF, BRJ
• Resource Availability
• RAI (Indicator)
• RAC (Confirm)

20
H.323 H.225 RAS Messages

• Gatekeeper Discovery
• GRQ, CCF, GRJ
• Terminal/Gateway Registration
• RRQ, RCF, RRJ
• Terminal/Gateway Registration
• URQ, UCF, URJ
• Disengage
• DRQ, DCF, DRJ

21
H.323 H.225 RAS Messages

• Status Queries
• IRQ info request
• IRR info request response
• IACK info request ACK
• INACK - info request NACK

22
H.323 Q.931 Messages

• Alerting
• Called user has been alerted, (phone is ringing)
• Call Proceeding
• Call has been established, no more call
  establishment information will be accepted
• Connect
• Acceptance of call by called party
• Setup
• Indicates H.323 party wants to setup a connection
  to called party

23
H.323 Q.931 Messages

• Release Complete
• H.225 (Q.931) call has been released, signaling
  channel is now open
• Status
• Sent when unknown call signaling message or a
  status inquiry message is received
• Status Inquiry
• Requests a calls status

24
H.323 H.245

• Establishes logical channels for transmission of
  H.323 data
• Negotiates
• channel usage
• master/slave configuration
• flow control
• Codec used
• H.245 ports
• 1024-5000 TCP in Cisco implementation

25
H.323 H.245 Messages

• Master/Slave Determination
• Determines which terminal will be master which
  will be slave in the call
• Terminal Capability Set
• Contains information on a terminals ability to
  send and receive multimedia streams
• Open Logical Channel
• Opens logical channel for transport of multimedia
  data
• Close Logic Channel
• Closes the logical channel between two endpoints

26
H.323 H.245 Messages

• Request Mode
• Receive terminal requests type of transportation
  from a transmit terminal
• Types of Modes
• Video
• Audio
• Data
• Encryption

27
H.323 H.245 Messages

• Send Terminal Capacity Set
• Instructs far-end terminal to send transmit and
  receive capabilities
• End Session Command
• Indicates the end of the H.245 session

28
H.323 Call Setup via Gatekeepers
29
Directory and Tier 1 Gatekeeper Call Setup
Directory Gatekeeper
3. LRQ
2. LRQ
Tier 1 Gatekeeper
Tier 1 Gatekeeper
4. LCF
1. ARQ
2. RIP
8. ACF
7. ARQ
5. ACF
6. Q.931 Call Setup
VoIP PBX
VoIP PBX
8. Q.931Call Proceed
H.245
RTP
IP Phone
IP Phone
30
Gatekeeper Peering and Redundancy
31
Codec ITU G.711

• G.711 is the international standard for encoding
  telephone audio on a 64 kbps channel. It is a
  pulse code modulation (PCM) scheme operating at a
  8 kHz sample rate, with 8 bits per sample, fully
  meeting ITU-T recommendations. The module is
  designed and tested on the TI TMS320C54x platform
  but can be ported to other DSP and RISC
  platforms, as well as MS Windows.
• Information from http//www.spiritcorp.com/

32
Codec ITU G.711(cont)

• Features
• Fully compliant with ITU-T G.711
• 64 kbit/s expander input rate
• 104 or 112 kbit/s expander output rate
• A-law or mu-law expander input
• Uniform PCM expander output
• 104 or 112 kbit/s compressor input rate
• 64 kbit/s compressor output rate
• Uniform PCM compressor input
• A-law or mu-law compressor output
• Selectable frame/buffer memory size according to
  the system needs
• Very simple application interface
• Compliant with TI's eXpressDSP standard. Code is
  reentrant, supports multithreading and dynamic
  memory allocation. At the same time allows direct
  (non-eXpressDSP) interface to enable static
  memory allocation
• Can be easily ported to any DSP or RISC platform

33
Codec ITU G.722.1

• G.722.1 is a low-bit-rate wideband coder, which
  codes speech at 24 kbps or 32 kbps. The quality
  at 32 kbps is the same as that of G.722 SB-ADPCM
  at 64 kbps. It uses a transform-coding scheme
  called Modulated Lapped Transform (MLT), with a
  20 ms frame size. The algorithmic delay is 40 ms
  (20 ms frame size 20 ms look-ahead).

34
Codec ITU G.722.1(cont)

• Supports all bit rates viz. 16/32 kbps at 16 khz
  sampling rate
• C callable API for initialization, encoding and
  decoding of speech data
• Supports Multi-channel capability
• Optimized implementation
• Bit Compliant with ITU-T test vectors
• Information found at http//www.ittiam.com/pages/p
  roducts/g722-1.htm

35
Codec ITU G.723.1

• G.723.1 is a speech compression algorithm
  standardized by ITU. G.723.1 has dual coding
  rates at 5.3 and 6.3 kbps. The vocoders process
  signals with 30 ms frames and have a 7.5 ms
  look-ahead and low distortion while passing DTMF
  tones through. The input/output of this algorithm
  is 16 bit linear PCM samples.
• Middle bit rate G.723.1 vocoder delivers one of
  the highest compression ratios of any of the
  current ITU standards without compromising speech
  quality. This vocoder can perform full duplex
  compression and decompression functions for
  multimedia, visual telephony, wireless telephony,
  and videoconferencing products.

36
Codec ITU G.723.1 (cont)

• Features
• Fully bit exact with ITU-T G.723.1
• 5.3 and 6.3 Kbps encoded bit stream rates
• Discontinuous transmission support (DTX) using
  Voice Activity Detection (VAD) and Comfort Noise
  Generation (CNG)
• Includes optional High Pass Filter and optional
  Post Filter
• Direct interface with PCM 8KHz sampled data. Both
  sample-by-sample and block based processing
  supported
• Very simple application interface
• Can be easily ported to any platform.

37
Codec ITU G.726

• ITU-T G.726 has speech compression and
  decompression at rates of 16, 24, 32 and 40 Kbps
  based on Adaptive Differential Pulse Code
  Modulation (ADPCM). It can be effectively used
  for speech compression in such applications as
  speech storing, digital circuit multiplication
  and telephony applications.

38
Codec ITU G.726 (cont)

• Features
• Fully bit exact with ITU-T G.726
• Sample-by-sample or block based analog input
• 16, 24, 32 or 40 Kbps bit stream rate
• A-law, mu-law and 14-bit uniform 8 kHz PCM
  input/output
• Direct interface with PCM 8KHz sampled data. Both
  sample-by-sample and block based processing
  supported
• Very simple application interface
• Can be easily ported to any DSP or RISC platform

39
Codec ITU G.728

• ITU-T recommendation G.728 Annex G is the
  fixed-point version of the coding of speech at
  16kbps using Low Delay Code Excited Linear
  Prediction (LD-CELP). It uses backward adaptation
  of predictors and gain to achieve an algorithmic
  delay of 0.625 ms. Under error-free transmission
  conditions the perceived quality of a 16 kbit/s
  LD-CELP codec is equivalent to that of a codec
  conforming to 32 kbit/s ADPCM. The codec is
  suitable for applications such as VoIP.

40
Codec ITU G.728(cont)

• Features
• API functions for initialization, encoding and
  decoding of speech data
• Supports Multi-channel operation
• Information from http//www.hellosoft.com

41
Codec ITU G.729

• ITU-T recommendation G.729 codec belongs to the
  Code-Excited Linear-Prediction coding (CELP)
  model speech coders and uses Conjugate-Structure
  Algebraic-Code-Excited Linear-Prediction
  (CS_ACELP) for coding speech signals at 8
  kbits/sec. The coder operates on speech frames of
  10 ms corresponding to 80 samples at a sampling
  rate of 8000 samples per second and the total
  algorithmic delay is 15 milliseconds. The encoder
  functionality includes Voice Activity Detection
  and Comfort Noise Generation (VAD/CNG) and the
  decoder is capable of accepting silence frames.
  G.729 provides near toll quality performance
  under clean channel conditions and is the default
  codec as prescribed by the Frame Relay Forum and
  is also suitable for voice over network (VoIP)
  applications.

42
Codec ITU G.729 (cont)

• Features
• C-callable API functions for initialization,
  encoding and decoding of speech data
• Voice Activity Detection and Comfort Noise
  Generation
• Supports Multi-channel operation and Reentrancy
• Code passes all test vectors specified by ITU-T
• Optimized implementation

43
Codec ITU G.729A

• ITU-T recommendation G.729 annex A (referred as
  G.729A) is the reduced complexity version of
  G.729 recommendation and operates at 8 kbits/sec.
  This version is developed mainly for multimedia
  simultaneous voice and data applications,
  although the use of the codec is not limited to
  these applications. This version is bit stream
  interoperable with the full version (G729). The
  coder operates on speech frames of 10 ms
  corresponding to 80 samples at a sampling rate of
  8000 samples per second and the total algorithmic
  delay is 15 milliseconds. The encoder
  functionality includes Voice Activity Detection
  and Comfort Noise Generation (VAD/CNG) and the
  decoder is capable of accepting silence frames.
  The performance of this codec may not be as good
  as the G729 in certain circumstances. The codec
  is suitable for voice over network (VoIP)
  applications

44
Codec ITU G.729A (cont)

• Features
• C-callable API functions for initialization,
  encoding and decoding of speech data
• Voice Activity Detection and Comfort Noise
  Generation
• Supports Multi-channel operation and Reentrancy
• Code passes all test vectors specified by ITU-T
• Optimized implementation

45
GIPS

• GIPS Enhanced G.711 - G.711 with GIPS developed
  enhancement providing superior packet loss
  robustness. GIPS Enhanced G.711 consists of the
  G.711 codec combined with an enhancement to
  provide packet loss robustness. Call setup is
  done with G.711 and the enhancement is detected
  and activated after call setup if both end points
  have the enhancement. The enhancement unit works
  similarly to an encryption method. The packets
  are transcoded to provide packet loss robustness
  instead of privacy. A SoundWare solution with
  GIPS Enhanced G.711 in combination with NetEQ
  provides a PSTN speech quality level at packet
  loss/delay rates up to 10. This is achieved
  without increasing the bit rate, and without
  significant increases in latency and complexity.
• Information provided by http//www.globalipsound.c
  om

46
GIPS (cont)

• GIPS Enhanced G.711QualityAt parity with PSTN,
  even under severe packet loss conditions
• SAMPLING RATE8 kHz
• BITRATEVariable, in average equal to
  G.711COMPLEXITY Very lowPACKET LOSS ROBUSTNESS
  Very highALGORITHMIC DELAYEqual to G.711
• VOICE ACTIVITY DETECTION Available
• COMPATIBILITYTransparent with G.711 at the
  end-points and is only activated if the both
  end-points are improved with a GIPS codecDTMF,
• FAX AND MODEM COMPATIBLEYes

47
Speech Codec Comparison

• Codec TypeRate Algorithmic Delay(ms)
• G.711 A-Law / µ-Law 64 0
• G.722 SB-ADPCM 64/ 56/ 48 0
• G.723.1/ AMP-MLQ/ACELP 6.3/ 5.3 37.5
• G.726 ADPCM 16/ 24/ 32/ 40 0
• G.727 Embedded ADPCM 16/ 24/ 32/ 40 0
• G.728 LD-CELP 16 lt 2
• G.729 CS-ACELP 8 15
• G.729 ACS-ACELP 8 15
• G.729 BCS-ACELP 8 15
• G.729 ABCS-ACELP 8 15

48
Notes on Table

• All codecs are voice-band and run at an 8kHz
  sampling rate, except for G.722 which has a 7kHz
  bandwidth and 16kHz sample rate
• These rates are nominal due to utilization of
  silence compression schemes
• G.711 and G.722 are provided free of charge with
  G.728 if required for H.320
• Algorithmic delay of "non-predictive" codecs is
  effectively zero
• Information from http//www.spa.com.au/faqs/codecs
  .html

49
Matching PSTN Quality
50
Better Than PSTN Quality
51
VOIP Codecs - bandwidth vs. Quality

• The tradeoffs
• How much do you need (quality)?
• How much can you afford?
• How much coding delay can you tolerate?
• Do you have special needs?

52
Issues with VoIP

• Firewalls
• NAT
• QoS
• Network Testing

53
VoIP Issues Firewalls

• A set of security mechanisms than an organization
  implements to prevent unsecured access from the
  outside world to its internal network.
• Typically work by blocking access of certain
  network protocols to specific ports.

54
VoIP Issues NAT

• Helps protect the intranet from exposure to
  unwanted traffic by providing one single external
  address to remote users.
• Translates local intranet addresses into an
  external address.
• Remote users connect to this external address to
  connect to the local user, without actually
  knowing its local address.

55
Issues with Firewalls and NAT

• H.323 requires the use of specific static ports
  for RAS messages, and a number of dynamic ports
  for RTP.
• SIP has one port (5060) for SIP messages, as well
  as dynamic ports for RTP.
• For these protocols to pass the firewall, the
  specific static and the range of dynamic ports
  must be opened for all traffic.

56
H.323 Ports used by Cisco equipment

• Source Call Manager Dest. Gatekeeper
• Description Type Destination Port
• H.225 RAS UDP 1719
• Source Call Manager Dest. Call Manager
• Description Type Destination Port
• H.225 Call Setup TCP 1720
• H.245 Call Control TCP 1024-5000

57
H.323 Ports used by Cisco equipment

• Source Gatekeeper Dest. Gatekeeper
• Description Type Destination Port
• H.225 RAS UDP 1719
• Source Terminal Dest. Call Manager
• Description Type Destination Port
• Skinny TCP 2000

58
H.323 Ports used by Cisco equipment

• Source Terminal Dest. Terminal
• Description Type Destination Port
• RTP UDP 1024-65535
• RTCP UDP 1024-65535

59
Port Usage

• Cisco Call Manager -- Call Control for IP Phone
• 2000 TCP
• Cisco Call Manager -- H.225 Signaling
• 1720 TCP
• Cisco IP Phone -- RTP Streaming
• 16384-32766 UDP (dynamically allocated)
• Cisco IP Phone TFTP, DNS
• 49152-53247 TCP (dynamically allocated)
• Cisco IP Phone Call Control
• 49152-53247 TCP (dynamically allocated)

60
Voice Over IP - the reasons that we have all
heard!

• Perception
• It is cheaper to run just one network.
• It is easier to integrate advanced technology
  when your phone is on the network (CTI).
• If you dont do it someone else will.
• Reality
• Convergence will occur some day so it is
  important that we build the required
  relationships now.

61
TCP/IP implementations

• Departmental VOIP PBX
• Centralized VOIP PBX
• Road warriors
• VOIP trunking
• Intranet
• Internet

62
VOIP decisions

• Power for VOIP phones
• E-911 mapping
• Which Codec to use

63
Parameters the impact VOIP- How much is to much?

• Packet loss
• Latency
• Jitter

64
QoS issues

• QoS has been added to the H.245 OLC packets to
  allow endpoints to set QoS parameters for the
  media streams, including RSVP parameters. H.323
  only communicates QoS information between H.323
  devices. Actual reservation and control of
  resources is outside the scope of the standard.

65
QoS options

• Prioritize by Application
• Prioritize by Address - many applications
• Create separate VLANs

66
(No Transcript)