Transcript MM_Introduction

Multimedia Services in the Internet
Dorgham Sisalem / Sven Ehlert
FhG Fokus
[email protected]
sip:[email protected]
1
Lecture Overview
• Introduction to the Course
– Goals
– Administrative Stuff
– Topics covered
• Basic Networking Principles Refresh
– PSTN
– IP Networking
– Routing
2
Goals
• Overview of multimedia service (SIP)
• Understanding of multimedia services in the Internet
• Understanding of the general pictures
– Transport protocols, signaling, traffic types, QoS
• Practical experience with protocols and applications
• Basic knowledge of the different involved protocols and
concepts
• We are not dealing with:
–
–
–
–
–
–
Audio and video compression
Web programming
Image processing or speech recognition
Audio and video hardware
MMS or video over GSM
Where to get the latest movies or how to copy a DVD
3
Structure
• Pre-requirements
– Good understanding of IP networking principles
• 2-Hour credit
• Written exam on 11.7
– There is no second exam date
– There are no other exam means (oral …)
• Office hours: After the lecture
• Contact:
– [email protected]
• Slides
– http://www.tkn.tu-berlin.de/curricula/ss06/vl-mkn/index.html
– Basically still valid but will be updated soon
4
References
•
•
•
•
•
•
www.ietf.org (RFCs and drafts)
www.iptel.org (SIP tutorial)
www.cs.columbia.edu/~hgs/internet Schulzrinne Overview
Stevens, „TCP/IP Illustarted, V1“ (basic protocols)
Ferguson, Huston, „Quality of Service“ (general QoS stuff)
Henry Sinnreich and Alan B. Johnston „Internet
Communication Using SIP: Delivering VoIP and Multimedia
Services with Session Initiation Protocol“
• Olivier Hersent, David Gurle, Jean-Pierre Petit,“IP Telephony“
• Huitema, „IPv6“
• Wikipedia
5
Acknowledgements
• Slides based on work of Henning Schulzrinne, Jim
Kurose, Michael Smirnov, Georg Carle, Jiri Kuthan,
Heikki Waris, Kevin Fall, Jim Chou, Thinh Nguyen,
Vishal Misra, Steve Deering, Geert Heijenk, Ofer
Hadar, John Floroiu, Nick McKeown, Eric D. Siegel,
Ibrahim Matta, Steven Low, Vincent Roca, Nitin H.
Vaidya, Charles Lang as well many other
anonymous contributers.
6
Topics: Introduction
• Introduction to Internet
– Very brief covering
• Difference between IP and PSTN
• Basic concepts
• Transport protocols: TCP, UDP, SCTP, RTP
– Why use UDP for VoIP and TCP for signaling?
– What are the benefits of SCTP
– What is the difference between RTP and RTCP
– You are expected to have visited the networking lecture of
Prof. Wolisz
7
Topics: VoIP
•
•
•
•
•
•
What is VoIP
Signaling
Addressing
Intelligent services
Deployment problems: NAT, emergency
Integration with PSTN
8
Topics: VoIP
What happens
during this
registration?
9
Topics: VoIP
What does this
address mean?
How do we find the
other side?
How do we call a
PSTN number?
What happens when
we press call?
10
VoIP in UMTS
•
•
•
•
•
What does IMS stand for?
Basic concepts of UMTS
What is the difference to normal VoIP?
How does it work?
Why a special version?
11
Problems of VoIP
• Why doesn’t VoIP work over my DLS link
– What are the problems of network address translators?
– How to deal with firewalls
• Regulatory issues
– How can I call the 110?
• QoS
– Is it true that VoIP has bad voice quality
12
Streaming
• How are resource described?
• What happens when we press play? (signaling)
• What does it mean when it says “buffering” or ran
out of buffer
• What protocols exist and how do they work?
13
Peer-To-Peer Networking
•
•
•
•
How do P-2-P solutions work?
What solutions exist?
What is Skype?
Basic concepts and approaches
14
Instant Messaging and Presence
•
•
•
•
What is presence and IM
Basic concepts and approaches
What solutions and technologies exist
What are the current standards
– ICQ, Yahoo, MSN, etc is NOT a standard
• Relation to VoIP
15
Group Communication
• What is the difference between broadcast and
multicast
• How does a conference bridge work
• What solution is best for which scenario?
16
Quality of Service (QoS) Control
• Why is the voice communication over the Internet
not understandable some times?
• What can we do at the end system to improve the
QoS?
• What can we do in the network to improve the QoS?
• Why can’t we find a network that deploys QoS
concepts
17
Public Switched Transmission Network
PSTN
18
Public Switched Transport Network
(PSTN)
•
•
•
•
•
Exists now for around 100 years
800 M Subscribers
Supports Voice and Data (Fax) services
Guaranteed bandwidth share
In one country only a few exist
– usually a big one controlling the whole network
• Cost of switching equipment high (A few millions for a carrier grade
switching component
• Signaling to session establishment and control based on SS7
• Hierarchical address structure (E.164)
International
Identity
1-3 digits
National
Identity
2-to-5
digits
User
Identity
11 to 5
digits
Subaddress
Up to 40 digits
19
PSTN Architecture in Germany
Fernnetz
AVSt
Auslandvermittlungsstelle
Ca. 50 HVSt
Hauptvermittlungsstelle
Ca. 550 KVSt
Knotenvermittlungsstelle
Ortsnetz
Ca. 500 OVSt
Ortvermittlungsstelle
Ca. 40 M Teilnehmer
20
Ref. Prof. Dr.-Ing. Habil. Lutz Winkler, FH Mittweida
Routing in PSTN
21
Ref. Prof. Dr.-Ing. Habil. Lutz Winkler, FH Mittweida
Switching in PSTN
Capacity 100
99 calls active
busy
22
Ref. Prof. Dr.-Ing. Habil. Lutz Winkler, FH Mittweida
Resource Sharing (TDM)
• Time division multiplexing (TDM)
–
–
–
–
10 kb/s
10 kb/s
10 kb/s
May under utilize channel with idle senders
Applicable only for a fixed number of flows
Requires precise timers
Resources are guaranteed
1 link, 30kb/s speed
Multiplexer
23
Intelligent Service in PSTN
24
Ref. Prof. Dr.-Ing. Habil. Lutz Winkler, FH Mittweida
Intelligent Service in PSTN
• Service switching point (SSP): A switch enhanced with logic
for identifying IN services
• Service Transfer Point (STP): Interface of the switch to the IN
environment
• Service Control Point (SCP): Control the execution of the
service
• Service Management System (SMS): Control and manage
the available services and provide the interface for adding
new ones
• Intelligent Peripheral: Additional components for providing
certain services such as announcements
• Feature Node: Execute services provided by private entities
(similar to SCP)
25
PSTN Summary
•
•
•
•
Guaranteed Quality of Service
Intelligence in the Network
Signaling and Media tightly coupled
Scalability and Extension difficulties
26
Introduction to the Internet
27
General Words
•
•
•
•
•
Since 21 Years with the same technology (TCP/IP)
Moved from 4 sites in 1968 to around 200 M hosts today
Flat addressing and routing architecture
Based on packet switching
(the) Internet: “collection of networks and routers that spans x
countries and uses the TCP/IP protocols to form a single,
cooperative virtual network”. (Comer)
• intranet: connection of different LANs within an organization
–
–
–
–
Private
may use leased lines
usually small, but possibly hundreds of routers
may be connected to the Internet (or not), often by firewall
28
Packet Switched Communication
End Users
End Users
Router
Data Packets (Voice, Video, Games, Signaling…)
29
What‘s a network?
• Host: Communication end point (PC, PDA, cell phone,
coffee machine ...)
• Link: carry bits from one place to another (or maybe to
many other places)
• Switch/gateway/router: move bits between links,
forming internetwork
– IP router receives a packet from one interface and sends it out
over another
1
2
1
2
30
What‘s a Protocol?
• Protocol: rules by which active network elements
communicate with each other
• protocols = “algorithms + data structures”
–
–
–
–
formats of messages exchanged
actions taken on receipt of messages
how to handle errors
hardware/operating-system independent
• real-life examples:
– rules for meetings
– conversational rules (interrupts, request for retransmission, ...)
31
Protocol Mechanisms (What Do
Protocols Do for a Living?)
• All or some of the following:
– addressing/naming: manage identifiers
– fragmentation: divide large message into smaller chunks to fit
lower layer
– resequencing: reorder out-of-sequence messages
– error control: detection and correction of errors and losses
• retransmission; forward error correction
– flow control: avoid flooding/overwhelming of slower receiver
– congestion control: avoid flooding of slower network
nodes/links
32
Architectural Requirements of the
Internet
• Generality
– Support ANY set of diverse applications,
• Heterogeneity
– Interconnect ANY set of network technologies
• Robustness
– More important than efficiency
• Extensibility
– More important than efficiency
• Scalability
33
End-to-End Principle
Foundation of the Internet architecture:
• Dumb network, smart end systems
– (Exact opposite of telephone network!)
• Dumb networks: require only least common service
– Datagram service: no connection state in routers
– Best effort: all packets treated equally.
– Can lose, duplicate, reorder packets.
• Smart hosts:
– Maintain state to enhance service for applications.
– “Fate-sharing”-- If a host crashes and loses communication state,
applications that are communicating share this fate.
34
Resource Sharing (Statistical)
• Statistical multiplexing
– Traffic is sent on demand, so channel is fully utilized if
there is traffic to send
– Any number of flows
5 kb/s
20 kb/s
5 kb/s
1 link, 30kb/s speed
Multiplexer
35
Resource Sharing (Statistical)
• Statistical multiplexing
– Resources are NOT guaranteed
– Need Mechanisms to prevent congestion and domination
5 kb/s
50 kb/s
5 kb/s
Multiplexer
1 links, 30kb/s speed,
50% Loss
36
Who runs the Internet?
• “nobody”
• standards: Internet Engineering Task Force
• names: Internet Corporation for Assigned Names and
Numbers (ICANN)
• numbers: IANA (Internet Assigned Numbers Authority)
• network: ISPs (Internet Service Providers), NAPs
(Network Access Points), DFN, . . .
• fibres: telephone companies (mostly)
• content: thousands of companies, universities, individuals,
...
37
How big is the Internet?
• Many measures:
–
–
–
–
–
–
networks (routed entities)
domains, host names (but: several names per host!)
directly (continuously) attached hosts (“ping’able”)
IP-connected hosts (SLIP, PPP)
firewalled hosts
e-mail reachable
• As of January 2006, over 1 billion people use the Internet
according to Internet World Stats
38
Host Count
39
What Networks are There?
• Access (ISP):
– Carry data from users
• Core
– Carry data from access
• Network peering points
– Connect networks together
• Some enterprises might be connected directly to
core networks
40
An Example Network
USER
Backbone
Local Loop Carrier
Point of Presence
41
RFCs and Drafts
• “Request for Comments”, since 1969
• most RFCs are not standards!
• Internet drafts: working documents, but often used for
prototypes
• edited, but not refereed
• numbered sequentially (Jan 2006: more than 4000)
• check the April 1 ones. . . (RFC 1149)
• ftp://ds.internic.net/rfc
42
TCP/IP Stack
TCP/IP
Application
Transport
Application
VoIP
Email ..
Transport
TCP, UDP,
SCTP
Network
Network
Link
Link
Link
Router
Host
Host
Network IP, IPv6
Ethernet
Cable,
UMTS
43
Internet Protocol
• Deliver an IP packet from host to host(s)
• Connectionless, unreliable
– No loss handling
– No flow or congestion control
VoIP
SMTP
ICMP
HTTP
FTP
RTP
DNS
UDP
TCP
IPv4/IPv6
PPP
Ethernet
GPRS
SONET
AALx
V.x
ATM
44
Internet Names
• Physical link address
– Ethernet, Token Ring, FDDI..
– Flat
• IP address
– Identify an interface
– Topological
• IP Name
– Identify the object to reach
– Hierarchical
45
IP Addresses
• Identify an interface not host:
– A host can have more than 1 address
• IP addresses are 32-bit numbers (4.3 billion of
them!)
• Divided into parts: (network prefix, host number)
• 4 decimal numbers, called “dotted quad”
• Each (decimal) number is one byte
– Example: 128.32.25.12
• Can generally be used in place of names
46
Special Addresses
• Private addresses: Only of meaning inside an intranet
– 172.16 through 172.31 16
– 192.168.0 through 192.168.255 256
• Loopback: 127.0.0.1 (local interface)
• Local broadcast: all 1 (receive by all members of link)
• Multicast:
– 224.0.0.0 239.255.255.255
– Do not describe a host or interface but a group of receivers
• Reserved: 240.0.0.0
255.255.255.255
47
Internet Packets
• A lot of headers describing the different layers
Phy
IP
UDP/
TCP
Body
48
IP Header
•
•
•
•
•
•
Version: 4 or 6
Header length: number of 32 bit words of header
Type of Service: delay, throughput, reliability, monetary
Total length: length of packet in bytes
Identification: identify packet
Flag:
– Do not fragment
– More fragments
•
•
•
•
•
•
Fragmentation offset: Distance from the first bit of the original packet
Time-to-Live: Avoid loops
Protocol: Which protocol is used (TCP, UDP, ICMP ..)
Header Checksum: Calculated over IP header
Source address: Address of sender
Destination address: Address of receiver
49
IPv6: Why move to another
protocol?
• Lack of IP addresses
– Support for nearly endless range of addresses
• Better handling of options
– Reduce complexity of IP header
• Better support for management and administration
– auto configuration and renumbering
– Support plug&play
• Higher Packet sizes (Jumbograms)
• Need for better support for mobile and secure communication
– Remove the need for network address translators
• Really?
50
IPv4 vs. IPv6 Header
•
•
•
•
•
•
•
•
14 fields, at least 20 octets
32 bit addresses
fragmented packet processing at every
hop
header checksum recalculation at every •
hop
variable Options field for extra
•
processing information
•
8 fields, fixed 40 octet size
128 bit addresses
fragmentation only in endpoints, or
lower layer
– Usage of Path MTU discovery
no checksums
– Already in lower layers
new 20 bit flow label field
options in Extension Headers
51
IP Names
host name (has IP address)
organization administering
host
Organization administering
subnames to left
organization type or country
Oxany.fokus.fhg.de
52
Getting From A to B
53
Getting from A to B
• Know name: need to know IP address
– Domain Name System (DNS)
• Know IP address: need to know the way
54
Getting From A to B
Name to IP Address
55
Domain Name System
• The Domain Name System (DNS) is a distributed
database that is used by TCP/IP applications to…
– map between hostnames and IP addresses,
– and to provide application routing information.
• Distributed database:
– No single site on the Internet “knows it all.”
– Each site maintains its own database and runs a server
that other systems on the Internet can query.
• DNS is the client/server protocol.
56
Domains
• Top level domains
– arpa domain
• Special domain for address-to-name mappings
– generic (organizational) domains
• 3-character domains (e.g. edu, com, org, …)
– Country (geographical) domains
• 2-character domains
• Found in ISO 3166
• Some countries form second-level domains
– e.g.: .ac.uk is for academic institutions in the United Kingdom.
– New generic top level domains (gTLD)
• .biz, .name, ,info ...
• Note: No single entity manages every node.
57
DNS hierarchical name space
unnamed root
top level domains
de
arpa
us
Maintained by DeNIC
com
edu
gov
wsu
eecs
gazoo
math
int
mil
net
org
•Node labels up to 63 characters.
•Root node has null label.
•Comparisons are case insensitive.
•Domain name formed as follows:
•start at node and work toward root
•use a “dot” to separate labels
58
Resolvers and Name Servers
• Applications (clients and servers) contact a DNS
server by calling functions in a library known as a
resolver.
– The resolver is accessed through the functions
gethostbyname() and gethostbyaddr().
– The resolver code is in a system library and is linked into
the application.
59
DNS Operation
• What does a server do when it does not have the
requested information?
– Every name server must know how to contact the root
name servers (via IP address).
– Name server contacts a root server
– Root servers know the name and IP address of all the
second-level domains
– Each names server caches information from recent
queries.
60
Practical
• nslookup
• http://www.internic.org
61
Routing Packets from A to B
62
Hierarchical PSTN Routing
030
040
050
060
63
Distributed IP Routing
193.175.135.21
Core
Access
PictureTel
Enterprise
Core
195.37.78.225
Access
Access
64
IP Routing
• How to get from A to B?
– Different paths are possible!!
– Neither A nor B know the best path in advance!!
• Goal: set routing tables for packet forwarding in hosts and
routers, typically based on some optimality criterion.
• Questions:
–
–
–
–
–
who determines entries?
based on what information (hops, delay, cost, ...) ?
how often does it change (hop vs. delay)?
where is routing information stored?
algorithm used to compute routes?
65
IP Routing: Goals
•
•
•
•
•
•
•
•
•
•
scalability
“safe” interconnection of different organizations
adopt quickly to changes in topology
avoid routing loops or at least terminate them quickly
self-healing, robust
Distributed: No central component to determine the path
efficient: can’t use 90% of bandwidth for routing info
multiple metrics (QOS, price, politics, ...) not yet
routes should be (near) “optimal”
can’t have all hosts/networks in single table hierarchical
66
IP Routing
• Every router needs to determine the next hop to which to send the data
• Routing database: one entry for every possible destination in the system:
–
–
–
–
–
–
Destination address: the IP address of the host or network;
Next hop: the first router along the route to the destination;
Interface: the physical network which must be used to reach the first hop
Metric: a number, indicating the distance to the destination;
Timer: the amount of time since the entry was last updated;
Flags and other internal information.
1
2
1
2
67
Intra-Domain Routing
• Set the routes inside an
autonomous system (AS)
– AS: a a collection of routers
and system administered
by one entity
– Has a AS number assigned
by IANA
• Different ASs might use
different intra-domain
routing schemes
• Changes in one AS do not
effect other domains
• AS connects to another AS
through one or more border
routers
Core
Access
Enterprise
Core
Access
Access
68