Voice over Internet Protocol VoIP

1

• Voice over Internet ProtocolVoIP
• 3. Issues with VoIP
• Barry Cheetham

2
Setting up a VoIP session
Look
Invite
Reply
Locationserver
Invite
Proxyserver
Barry
All messages could be sent by TCP
Alvaro

• B want to call A but has no idea where he is or
  whether he is online.
• A registers with location server
• B invites A via proxy server that looks up As IP
  address.

3
Maintaining a VoIP session
Look
Accept
Reply
Locationserver
Accept
Proxyserver
Voice
Barry
Alvaro
A can accept, B acknowledge, then interchange
packets via server. Instead establish RTP/RTCP
link between B A peer-to-peer. Negotiate
bit-rate by monitoring congestion.
4
Proxy location servers

• Location server like DNS (in 1st lecture)
• Proxy server is intermediate software that acts
  as both server client to make requests on
  behalf of other clients.
• Useful for routing as gate-keeper for
  enforcing policy such as admission control,
  preventing congestion, efficient access to
  location server, etc.
• Since UDP sometimes blocked by firewalls, a PS
  may use tunnelling by encapsulating UDP packets
  within TCP.
• Also gateway between IP network networks.
• Not much encryption used in VoIP yet.

5
QoS in VoIP

• Jitter-buffer at each receiver allows for
  variation in delay.
• Assume 1-way delay varies between 0.02 0.12 s,
• Jitter-buffer of size 0.1 s to avoid data
  under-flow
• i.e running out of data while waiting for delayed
  packet.
• Receiver has no way of knowing actual 1-way
  delay.
• Delay variation from 0.05 to 0.15 gives same
  jitter.
• Packet arriving too late to avoid underflow is
  lost.
• With wired networks, most lost packets are just
  late.
• Packets rarely undelivered or damaged.
• Not true for wireless/mobile networks.

6
Jitter-buffer size

• Receivers jitter-buffer introduces delay
  affects number of packets lost due to excessive
  delay.
• Increasing buffer size decreases no. of lost
  packets in a wired VoIP link at expense of
  increasing delay.
• Round trip delay gt 300 ms makes conversation
  difficult.
• Limits delay that can be introduced at each
  receiver.
• Assume propagation delay is 30 ms, payload is
  160 bytes (20ms)
• 2 50 ms used up
• Capacity for 100 ms buffers at each end of 2-way
  link.
• Reservoir of 800 bytes or 5 packets.

7
Question for you

• Having set up such a VoIP link between Manch
  NY,
• how could you improve it if RTCP reports zero
  lost packets at NY end 10 at the Manch end?
• Answer Increase buffer at Manch end to decrease
  no. of lost packets at expense of increasing
  round trip delay.
• Decrease buffer at NY to decrease delay at
  expense of increasing lost packets.
• No discernable change in delay.
• Voice quality improves at Manch gets a bit
  worse at NY.

8
H323 application layer protocol suite

• Protocols to set up, maintain tear down VoIP
  calls. .
• Once call set up, voice packets sent by RTP/RTCP.
  
• 64 kb/s speech samples conveyed in blocks of 240
  samples (30 ms).
• Simultaneous video signals may be sent also.
• Use of RTP means no QoS guarantees.
• Use of 64 kb/s places high demand on IP networks
• Can use speech compression to reduce bit-rate.
• H323 negotiates suitable compression based on
  prevailing network condition.
• Has gatekeepers gateways (mentioned
  earlier).

9
SIP application layer protocol

• Session Initialisation Protocol (Ports 5060
  5061)
• Defined by IETF as simpler alternative to H323
• Addresses are URLs, e.g. sipbarry_at_man.ac.uk.
• Uses RTP RTCP (like H323) for voice streams.
• Adapts to non-guaranteed QoS like H323, has
  proxy servers acting as gatekeepers, etc.
• Win Live Messenger based on SIP.
• More than 100 different VoIP systems most use SIP

10
Speech digitisation standards
G711 widely used for VoIP lower bit/rate coders
available. Many issues summarised by table below
11
Mobile VoIP Voice over wi-fi

• Application of VoIP to battery powered wireless
  enabled mobile devices including handsets.
• Since WLANs support IP, UDP TCP, they can
  support VoWiFi.
• H323 or SIP protocols will work on wi-fi
  connected devices.
• Wireless QoS quite different from wired
  bit-errors more frequent, bandwidth more limited,
  congestion more likely.
• Each packet more expensive to send.
• Employ forward error correction (FEC) next
  years course.
• Lets just illustrate these issues

12
Wi-fi packet lengths (Phy layer)
13
Illustration of Voice over wi-fi

• Phy-layer header takes ?200 ?s in addition to
  payload.
• 1000 bytes of text takes ?2 ms to transmit at 11
  Mb/s, so overhead is about 10,
• 100 bytes sent in 0.2ms, phy layer overhead
  becomes 100 .
• For 30 bytes, framing overhead approaches 300.
  
• Voice over wi-fi begins to appear rather
  inefficient.
• Increasing packet size not an option with
  interactive VoIP.
• If retransmissions become necessary because of
  collisions radio noise, inefficiency becomes
  worse.
• Thats all on VoWifi

14
Packet loss concealment strategies

• PLC uses predictability in speech waveforms.
• Allows guesses about what is likely to come next.
  
• Works well for voiced sounds quasi-periodic.

t

• If dashed part lost, similarity with what came
  before makes it possible to produce a
  reasonable replacement.

15
Simple frame repetition

• Obvious idea is just to repeat the previous
  frame,
• Works perfectly in previous example.
• Much better than zero stuffing.
• Simple packet repetition will not always work so
  well,
• Wave-shape its periodicity will be changing
• Any discontinuity as reconstructed frame joins
  onto next frame will produce a nasty click.
• Need something a bit more sophisticated
• Provided by Appendix 1 to the ITU-G711 standard.
  
• Demo (30 PL) Zero stuffing
  Repetition

16
Some finer points of real time

• Why not forget about the missing packet go on
  with next one?
• Just miss it out?. OK for a file of music?
• With a video sound track, you may lose lip-sync.
  
• Disastrous for telephony because of real time
  requirement.
• No. of samples sent must be no. of samples
  received
• Otherwise receivers jitter-buffer runs out
  generates a click.
• Expect small differences in sampling rates at
  transmitter receiver

17
IntServ Diffserv

• Reservation of transmission capacity priority for
  certain traffic.
• VoIP widely used over wired networks with
  reserved capacity .
• IETFs IntServ architecture proposed for giving
  guaranteed QoS to particular traffic streams by
  reserving link capacity between routers.
• IETFs DiffServ architecture proposed for
  allowing certain categories of traffic, such as
  VoIP, to be prioritised to make it more likely
  (though not certain) to achieve a desired QoS.

18
Some fallacies about distributed computing

• Interconnections are
• Reliable
• Secure
• Homogenous (same type of links throughout)
• Unchanging in topology
• So fast that latency (delay) is negligible
• Essentially unlimited in bandwidth.
• Available at zero transport cost (is VoIP free?)
• Set up with just one administrator
• Linking devices with fully synchronised clocks
• i.e. each having a universal measure of
  time (144511.001)
• Many of these fallacies are revealed by VoIP

19
Transparency

• A distributed system for VoIP should
• Accommodate differences in data representation
• Not worry about where resources are located
• Allow resources to move while in use
• Allow resources to be replicated.
• Allow processing to be distributed among hosts
• Be robust to failure of components one failure
  should not be catastrophic

20
Conclusions learning outcomes

• Interactive VoIP requires regular transmission of
  packets with round trip latency limits.
• End to end delay not easily measurable not
  needed
• H323 SIP for setting up maintaining ending
  calls.
• Fire forget transport layer protocols RTP
  RTCP.
• Bit-rate compression negotiated .
• No QoS guarantees jitter lost packets occur.
• VoWLAN or VoWiFi works in principle but less
  efficient
• Wired wireless networks have different QoS.
• Principles of PLC algorithms discussed.
• Intserve diffserve for QoS.
• Some fundamental issues illustrated by VoIP

21
A final question (for today)

• An end-to-end VoIP over wi-fi system uses 50 ms
  G711 speech frames with zero stuffing PLC.
• It delivers intelligible but fairly poor quality
  speech, with round trip delay of 200 ms and
  frequently crashes due to network congestion.
• How could you improve its performance?
• What steps would be appropriate if it were a
  wired link?