14: Signalling Protocols
Mark Handley
H.323
ITU protocol suite for audio/video conferencing over “networks that do
not provide guaranteed quality of service”.
H.225.0 layer
Source: microsoft.com
1
H.323 stack
H.323 User Interface
Multimedia Applications,
Data
Media Control Terminal Control and Management
Applications
Audio Video
Codecs Codecs
G.711 H.261
G.723.1 H.263
G.729 H.264 H.225.0
.. .. H.225.0
V.150 T.120 T.38 RTCP Call H.245
RAS
Signaling
RTP
UDP TCP TCP/UDP UDP TCP/UDP TCP UDP
IP
Source: packetizer.com
H.323 Protocol Components
A/V codecs (H.261, H.263, G711, G.723).
H.225.0 transport
Uses RTP/RTCP for audio/video packets
Includes RAS (Registration, Admission and Status) signalling
channel for communication with the gatekeeper.
Q.931
“ISDN user-network interface layer 3 specification for basic call
control”
H.245 call control
Negotiation of codec, bit rate, frame rate, etc.
T.120 data communications.
2
H.323 System Components
Terminal
videophone, MS netmeeting software, etc
Gateway
Gatekeeper
MCU (multipoint control unit).
Gateways
Optional element in an H.323 conference.
Not usually needed for pure H.323 to H.323 calls.
Principle role is translation function between H.323
conferencing endpoints and other terminal types. Eg:
Establishing links with analog PSTN terminals.
Establishing links with remote H.320-compliant terminals
over ISDN-based switched-circuit networks.
Establishing links with remote H.324-compliant terminals
over PSTN networks.
3
Gatekeepers
Optional component used for admission control and
address resolution. Acts as the central point for all calls
within its zone and provides call control services to
registered endpoints.
May allow calls to be placed directly between endpoints
May route the call signaling through itself to perform
functions such as follow-me/find-me, forward on busy,
etc. Service providers can also use this to bill for calls
placed through their network.
Can be used to limit the total conferencing bandwidth to
some fraction of the total available.
Multipoint Control Units (MCU)
Responsible for managing multipoint decentralised
conferences.
Three or more endpoints.
MCU contains:
Multipoint Controller (MC) that manages
MCU
the call signaling, and handles H.245
negotiations between all terminals to
determine common capabilities for A/V
processing.
Multipoint Processors (MPs) to handle
audio and video mixing, switching, or other centralised
media processing. MP is an optional control
component of the MCU. data
4
High level call flow
3. Collect replies to
previous query
4. Grant GK GK 7. Grant
permission to 2. Try to resolve the address permission
place call of the called party
6. Request
1. Request permission to
Permission to accept call
place call 5. Attempt to establish
the call
GW GW
8. Indicate connection
establishment
H.323 Call Progress
H.323 has several different
ways a call can progress
Direct mode vs Routed
Mode
Regular call vs Fast
start call (in H.323v2)
TCS = Terminal Capability Set
MSD = Master/Slave Determination
DRQ = Disconnect Request
[SourceL H.323 forum]
5
H.323 Usage
Microsoft NetMeeting (obsolete)
Lots of commercial videoconferencing equipment
Eg: Polycom
Some IP phones (including some of Cisco’s)
IETF Multimedia Protocol Suite
RTP/RTCP for data flow, A/V sync, and reception feedback.
SDP (Session Description Protocol) for describing a multimedia
session
Also used for negotiation of session parameters.
SAP (Session Announcement Protocol) for broadcast-style
announcement of multicast sessions.
SIP (Session Initiation Protocol) for setting up an reconfiguring
multimedia calls.
Telephony-like signalling.
Instant Messaging.
RTSP (Real Time Streaming Protocol) for remote control of VCR-
style functionality.
6
Session Description Protocol (SDP)
SDP is a standard way to describe multimedia sessions.
These descriptions can then be used in different contexts:
Session Announcements using SAP
Session Invitations using SIP
RTSP stream descriptions
H.332 announcements
PINT (PSTN/Internet IN feature mapping)
Advanced Television Enhancement Forum (!)
SDP was really only designed for SAP - the other uses stretch it a
little beyond its design space.
Sometimes this shows.
SDP
SDP is a text-based description format.
It is extensible through attributes (which don't have to be
registered) and by several other namespaces that are
registered with IANA.
It was not intended for content negotiation.
SIP can use it for this purpose, but it's not as elegant as
if SDP had been designed for this.
7
SDP
It was intended for announcing the existence of sessions. It conveys:
Information to allow you to choose whether to join the session.
Session timing information
Information to inform you of the resources required to participate.
Sufficient information to allow you to join the session
protocols and codec formats
multicast addresses and ports
encryption keys
Information that RTP needs passed out-of-band.
SDP: Example
8
SIP: Session Initiation Protocol
Original spec:
RFC 2543
Updated specs:
RFC 3261 (main spec)
RFC 3262 (provisional response reliability)
RFC 3263 (locating SIP servers)
RFC 3264 (offer/answer use of SDP)
RFC 3265 (specific event notification)
SIP: Early History
Early session initiation protocols:
ivsd (Turletti, INRIA)
mmcc (Schooler, ISI)
These led eventually to SIPv1 (Handley and Schooler) which was
intended to initiate loosely coupled sessions.
SIP stopped when the session started
UDP based
Henning Schulzrinne designed SCIP using RTSP as a basis around the
same time.
TCP based
Continued during the call
Eventually we merged SIPv1 and SCIP into SIPv2
best features of both, UDP and TCP
Note: H.323 was also being drafted at the same time
9
SIP: Aims
Both SIPv1 and SCIP were trying to allow user mobility.
Few people used mmcc because people move around
and change hosts.
In the merged draft, this became a key issue.
Goal was to be able to support invitations to public
multiparty multicast sessions or to private sessions.
SIP may or may not continue to be involved after joining
the session.
SIP: User Location
There are two basic ways to do user location:
Have a distributed directory.
Lookup user's location in directory.
Address a call to that location.
Lookup during call routing
Lookup a well-known address for the user.
Route the call there. Let them do the lookup of user location.
Either relay or redirect the call.
Multiple lookups can occur if required.
In the Internet, heterogeneity is key. Uniform distributed directories such as X500 have
failed to be deployed.
Lookup during call routing allows heterogeneity of user-location mechanisms.
Improved security.
10
SIP: Relaying a Call
SIP: Redirecting
a Call
11
SIP Proxies
SIP proxies can use any reasonable search algorithm
Send requests in parallel
Send requests sequentially
Normally only a proxy close to the callee can decide on an
appropriate search strategy.
SIP specifies only the rules that proxies must use to combine
responses when multiple requests are made in parallel.
A standard way to specify proxy call processing rules is desirable, but
SIP itself doesn't care how the processing is performed.
SIP: User Location Servers
12
SIP: User Location Servers
SIP doesn't need a separate user location server in many
circumstances:
SIP Normal protocol operation
13
SIP Syntax
SIP is a text based protocol, similar in syntax to HTTP and
RTSP.
Messages can be conveyed over UDP or TCP.
SIP provides its own reliability over UDP.
UDP is prefered - it gives more control over message
timing, and requires less state in proxies.
TCP is allowed for legacy firewall traversal but in time
we hope firewalls themselves will support SIP.
Typically SIP carries an SDP session description as a
payload to describe the session being initiated.
SIP Request (sent to north.east.isi.edu)
INVITE sip:
[email protected] SIP/2.0
Via: SIP/2.0/UDP east.isi.edu <-- second relay
Via: SIP/2.0/UDP isi.edu <-- first relay
Via: SIP/2.0/UDP chopin.cs.caltech.edu <-- originating host
To: sip:
[email protected] <-- original destination
From:
[email protected] <-- sender
CSeq: 1 <-- command seq. no.
Content-Type: application/sdp
Content-Length: 214
________________________________________________ start of payload _____
v=0
o=eve 987329833 983264598 IN IP4 128.32.83.24
s=Quick Call
...
14
SIP Response (sent to east.isi.edu)
SIP/2.0 200 OK
Via: SIP/2.0/UDP isi.edu Via field for east removed already
Via: SIP/2.0/UDP chopin.cs.caltech.edu
To: sip:[email protected] Refers to request "to", "from"
From: sip:[email protected] not message to and from.
Location: sip:[email protected];tag=76fa98c80aba81
CSeq: 1
Content-Type: application/sdp
Content-Length: 214
v=0
o=eve 987329833 983264598 IN IP4 128.32.83.24
s=Quick Call
...
SIP Usage
Almost all IP phones
Microsoft Windows
Messenger
Apple iChatAV
AT&T, MCI VoIP service
Sprint PCS cellphone
(walkie-talkie service)
3G cellular: IP Multimedia
Call Control
many more...
15
RTSP: Real-Time Stream (Control) Protocol
RTSP provides a way to set up and control multimedia
streams from a media server.
Essentially RTSP is the remote control for a network-VCR.
RTSP is in the same "protocol family" as SIP and HTTP:
text based
MIME-format messages
HTTP-like syntax
shared response codes.
RTSP functionality
Setup a connection and exchange stream transport
information
Describe the session
Play the session
from specified start times
forwards, backwards
at different speeds and data rates
Record a session
Pause playback or recording
16
RTSP example
Client to HTTP Server:
GET /twister.sdp HTTP/1.1
Host: www.example.com
Accept: application/sdp
HTTP Server to Client:
HTTP/1.0 200 OK
Content-Type: application/sdp
v=0
o=- 2890844526 2890842807 IN IP4 192.16.24.202
s=RTSP Session
m=audio 0 RTP/AVP 0
a=control:rtsp://audio.example.com/twister/audio.en
m=video 0 RTP/AVP 31
a=control:rtsp://video.example.com/twister/video
RTSP example
Client to Audio Server:
SETUP rtsp://audio.example.com/twister/audio.en RTSP/1.0
CSeq: 1
Transport: RTP/AVP/UDP;unicast;client_port=3056-3057
Audio Server to Client:
RTSP/1.0 200 OK
CSeq: 1
Session: 12345678
Transport: RTP/AVP/UDP;unicast;client_port=3056-3057;
server_port=5000-5001
17
RTSP example
Client to Video Server:
SETUP rtsp://video.example.com/twister/video RTSP/1.0
CSeq: 1
Transport: RTP/AVP/UDP;unicast;client_port=3058-3059
Video Server to Client:
RTSP/1.0 200 OK
CSeq: 1
Session: 23456789
Transport: RTP/AVP/UDP;unicast;client_port=3058-3059;
server_port=5002-5003
RTSP example
Client to Video Server:
PLAY rtsp://video.example.com/twister/video RTSP/1.0
CSeq: 2
Session: 23456789
Range: smpte=0:10:00-
Video Server to Client:
RTSP/1.0 200 OK
CSeq: 2
Session: 23456789
Range: smpte=0:10:00-0:20:00
RTP-Info:
url=rtsp://video.example.com/twister/video;
seq=12312232;rtptime=78712811
18
RTSP example
Client to Audio Server:
TEARDOWN rtsp://audio.example.com/twister/audio.en RTSP/1.0
CSeq: 3
Session: 12345678
Audio Server to Client:
RTSP/1.0 200 OK
CSeq: 3
Client to Video Server:
TEARDOWN rtsp://video.example.com/twister/video RTSP/1.0
CSeq: 3
Session: 23456789
Video Server to Client:
RTSP/1.0 200 OK
CSeq: 3
RTSP Usage
RealPlayer (and Helix open-source version)
Microsoft Windows Media 9, 10
Also supports Microsoft’s proprietary mms to talk to
older clients.
Apple Quicktime Player
3G cellular video streaming.
19
References
ITU-T Recommendation H.323 “Packet-based multimedia
communications systems” http://www.itu.int
RFC2327, “SDP: Session Description Protocol” M.
Handley, V. Jacobson
RFC3261, “SIP: Session Initiation Protocol”, J.
Rosenberg, H. Schulzrinne, G. Camarillo, A. Johnston, J.
Peterson, R. Sparks, M. Handley, E. Schooler
RFC2326, “Real Time Streaming Protocol (RTSP)” H.
Schulzrinne, A. Rao, R. Lanphier
20
