Transcript Chapter 2 Protocol Architecture - Department of Computer Science

Data Communications and
Networking
Chapter 2
Protocol Architecture:
TCP/IP model and OSI Model
Reading:
Book Chapter 2
Data and Computer Communications, 8th edition
By William Stallings
1
Outline
• Section 1: The need for a protocol architecture
• Section 2: TCP/IP protocol architecture
• Section 3: OSI model
• Learning outcomes:
—Describe the network protocol architectures and
reference models
2
Section 1
The Need for a Protocol Architecture
3
What is a protocol?
• From dictionary:
—A general definition:
the accepted or established code of procedure or
behavior in any group, organization, or situation
—For computing:
set of rules governing the exchange or transmission
of data electronically between devices
4
Protocol
• To perform a task, the involved parties usually follow a
common protocol designed for this task
— The protocol is just a set of rules or conventions
— Different tasks use different protocols
• Some key features of a protocol
— Syntax
• Concerns the format of the data blocks
— Semantics
• Includes control information for coordination and error handling
— Timing
• Includes speed matching and sequencing
5
Examples of Protocol
• Two protocols will be studied in your first lab
— HTTP is the protocol used for browsing website
• Web browser <----> Web server
• http://www.youtube.com/watch?v=1IQFjTnDozo&feature=related
— SMTP is the protocol used for sending emails
• Email client software < ---- > SMTP server
• A new application usually uses a new protocol, e.g. BitTorrent protocol:
http://en.wikipedia.org/wiki/BitTorrent_(protocol)
• TCP is a protocol used by Internet hosts for reliable data transmission
— Recall that Internet packets could be lost inside the Internet
• IEEE 802.11 is the protocol used in WiFi
— How to coordinate a number of wireless devices within a WiFi network?
6
Need For Protocol Architecture
• There are lots of network applications
• Building each application from scratch is very
time-consuming and challenging
—What commands should be supported?
—How to respond to each command?
—How to identify the two peer applications?
• Each computer may run multiple applications!
—How
—How
—How
—How
to
to
to
to
identify the two computers?
convert the data into bit stream?
convert the bit stream into signals?
detect and handle data loss and data error?
• The network is not perfect!
—Etc.
7
(Cont.)
• Most of the network apps share some common
modules
Application A
Application B
Application C
Module A
Module B
Module C
Module T
Module T
Module T
Module I
Module I
Module I
Module N
Module N
Module N
Module P
Module P
Module P
Common
modules
8
(Cont.)
• A complex task is broken into subtasks: modular
design
— Each subtask is implemented separately as a layer, arranged in
a vertical stack
• Each layer performs a related subset of the functions
required to communicate with another system.
— It relies on the next lower layer to perform more primitive
functions and to conceal the details of those functions.
— It provides services to the next higher layer.
— Layers should be defined so that changes in one layer do not
require changes in other layers.
• So, instead of using a single complex protocol, it’s more
flexible to implement a stack of protocols!
— Reduce the design and development workload significantly!
9
Vertical Stack
Application A
Application B
Application C
Application developers
Module A
Module B
Module C
Module T
Operating System
Module I
Module N
Module P
Hardware
can be replaced by
other modules
Module W
10
Example
• E.g., A can only speak Chinese, B can only
speak Spanish, how can A communicate with B?
—A finds a translator C, who can speak Chinese and
English
—B finds a translator D, who can speak Spanish and
English
• Two layers:
—Higher layer: A and B
—Lower layer: C and D
11
A Two-layer example
Higher A
Layer
How to communicate?
messages
in Chinese
Lower
Layer
B
messages
in Spanish
messages in English
D
C
Lower layer provides services to the next higher layer.
12
Section 2
TCP/IP Protocol Architecture
13
Protocol Architecture
• Tasks of communications are broken up into modules
— Each module (or layer) can have its own protocol
• In very general terms, communications can be said to
involve three components: applications, computers, and
networks.
• For example, file transfer could use three modules (or
layers)
— File transfer application
— Communications service module
— Network access module
• The stack of protocols is called “Protocol Stack”
— Or Protocol Architecture
14
TCP/IP Protocol Architecture
• Developed by the US Defense Advanced Research Project Agency
(DARPA) for its packet switched network (ARPANET)
• Used by the global Internet
• It consists of a large collection of protocols that have been issued
as Internet standard by the Internet Architecture Board (IAB).
— Check http://www.ietf.org/rfc/rfc2026.txt
• The TCP/IP protocol architecture organizes the communication task
into five relatively independent layers:
—
—
—
—
—
Layer 5: Application layer
Layer 4: Transport layer, or Host to host (TCP belongs to this layer)
Layer 3: Internet layer, or Network layer (IP belongs to this layer)
Layer 2: Network access layer, or Link layer
Layer 1: Physical layer
• Remark: Each layer can have lots of different protocols!
15
Example
• World Wide Web
—Replies on the HTTP protocol
Web browser
Web server
Layer 5
HTTP
HTTP
Layer 4
TCP
TCP
Layer 3
IP
IP
Layer 2
IEEE 802.11
IEEE 802.3
Layer 1
IEEE
802.11g
IEEE 802.3
1000BASE-SX
16
Benefit of layering
• The most challenging problem: how to provide a
reliable data transfer service on top of an
unreliable data network?
—This problem is so important that today’s Operating
Systems all provide such reliable service.
• The burden of network application developers
has been reduced significantly!
—Hence the application developers can simply focus on
the application layer issues.
—We can easily develop thousands of network
applications.
17
Physical Layer
• The job of physical layer is to send individual bits from
one node to a directly connected node.
— Address the communications issue
• The physical layer also covers the physical interface
between a data transmission device (e.g., computer)
and a transmission medium or network.
• It is concerned with specifying characteristics of the
transmission medium, the nature of the signals, the data
rate, and related matters.
• Examples:
— Ethernet has many physical layer protocols: one for twisted-pair
copper wire, another one for coaxial cable, some others for
optical fiber, and so on.
— IEEE 802.11b, 802.11a, 802.11g are different physical protocols
for Wireless LAN, each with different transmission capabilities.
18
Network Access Layer
• It is also called “Link layer”.
• It solves the problem of exchanging data between two or more
directly connected devices (computers, switches, routers, etc).
The link layer packets are usually referred to as frames.
— Point-to-point communications
— Multiple access communications
• The services provided by network access layer depend on the
specific link-layer protocol that is employed over the link.
• Issues to be addressed:
— Reliability (error detection and error correction)
— Priority (some data may be more important than others)
— Addressing (for multiple access)
• Examples of link layer protocols:
— Point-to-point Protocol (PPP)
— Ethernet: IEEE 802.3
— Wireless LAN: IEEE 802.11
Multiple access
Point-to-point
19
Internet Layer (IP)
• IP layer is responsible for moving network layer packets
known as datagrams from one host to another.
— Main challenge: how to find a path from the source to the
destination?
• Two principal components:
— IP protocol: defines the format of the datagram, and how the
end systems and routers act on the datagram.
[http://www.ietf.org/rfc/rfc0791.txt ]
— Routing protocols: determine the routes that datagrams take
between sources and destinations. The Internet uses many
routing protocols.
• The Internet layer routes a datagram through a series of
packet switches called routers between source and
destination.
20
Postal System
To: Mr. Jacky Chan
No. 123, XX Road,
XXX, USA
21
Inside Postal System
Central Post Office (New York)
The delivery of
your package
depends on the
postal address.
Central Post Office (Beijing)
Central Post Office (Hong Kong)
Local Post Office (Kowloon Tong)
Local Post Office (Tseung Kwan O)
22
What are the main functions of
Internet layer?
•
•
Ultimate purpose: to provide a best-effort, connectionless, or datagram service
between end systems
— A host can send packets to any other host on the Internet
— Best-effort: No guarantee of packet delivery. Packets could be lost, contain errors, or out-oforder.
How to achieve the purpose?
— Addressing:
•
•
Every host needs to have a unique IP address
E.g., 158.182.6.41 is the IP address of our cslinux1 server.
•
It’s not possible to connect any two hosts directly, therefore we need some “switches” to help. These
“switches” are called IP Routers.
— Networking
•
E.g., if host A wants to send a packet to host B:
1. If A and B are in the same network:
— A can send the packet to B directly, without going through a router
2. If A and B are in different networks (very likely!):
— Condition: A must be connected to a router R1, B must be connected to a router R2, R1 may not be
connected to R2 directly, but R1 must be able to talk to R2 (through other routers)
— A first sends the packet to R1
— R1 helps to send the packet to R2 (this is a core part of the Internet layer) by checking the
destination IP address (i.e., B’s IP address) contained in the IP packet
— R2 sends the packet to B
23
IP Operation
LLC and MAC
are sub-layers of
Layer 2.
24
IP Address
• IPv4 address: 32-bit
• Dotted decimal notation
— 192.5.48.3 <--> 11000000 00000101 00110000 00000011
• Each IP packet includes the source IP address and the
destination IP address
— So that the routers know how to forward
• Normally, we can think of a 32-bit IP address as having
an network portion and local portion, where the network
portion identifies a network, and the local portion
identifies a host in that network.
— E.g., in IP address 158.182.7.15
• 158.182.7 identifies a network of COMP department
• 15 identifies a PC in the network
25
Routers
• Routers are responsible for receiving and forwarding packets
through the interconnected set of networks.
— Each router is connected with a number of other routers.
— Routers use buffer to store the incoming packets.
— Incoming packets could be dropped if the buffer is too full.
• Each router makes routing decision based on a routing table.
— Where to send out each IP packet?
— Solution: make a search in the routing table based on the destination
IP address
• How to build up the routing table?
— Static method: input by network administrators
— Automatic method: by using routing protocols
• Routers exchange routing information using a special routing protocol to
build routing tables.
• We will learn the principles of routing protocols in this subject.
26
Example of a Routing Table
21.x.x.x
F
18.x.x.x
B
C
A
D
E
145.26.x.x
G
Routing table at router A:
Network ID
Next Hop
18.x.x.x
B
21.x.x.x
C
145.26.x.x
D
212.134.7.x
E
212.134.7.x
27
Transport Layer
• How do software developers write network application programs?
— They make use of the services provided by Transport Layer, which is
normally part of the Operating System.
— In fact, TCP/IP is indispensable for all the contemporary Operating
Systems (even for small devices like PDA or smart phone!).
• Transport layer transports application-layer messages between the
client application and the server application.
— Today’s multi-task computer can have a number of simultaneous
applications. One task of Transport layer is to differentiate the data
from/to different applications.
— Another possible task of transport layer is to provide reliability.
• Make the life of application developers much easier!
• How does Transport layer send out the message?
— Divide the message into packets
— Use the Internet Layer to send out each packet
28
Overview of TCP and UDP
• Two most commonly used transport protocols: TCP and UDP
• TCP:Transmission Control Protocol [http://www.ietf.org/rfc/rfc793.txt ]
— connection-oriented
• Temporary logical association between entities in different systems
— Provides reliable data transmission service
— Includes source and destination port numbers
• Identify respective applications
• A connection refers to a pair of ports
— The data are organized into TCP segments
• UDP: User Datagram Protocol [ http://www.ietf.org/rfc/rfc768.txt ]
— Lacks of reliability
• No guaranteed data delivery
• No preservation of sequence
• No protection against duplication
— With minimum overhead
— Includes source and destination port
• An Internet application needs to choose either TCP or UDP.
29
UDP
• User Datagram Protocol
—Defined in RFC 768
—http://www.ietf.org/rfc/rfc768.txt
• The UDP provides an unreliable connectionless
delivery service using IP to transport messages
between machines.
—Because IP is unreliable
—An application program the uses UDP needs to
handle the problem of reliability, including message
loss, duplication, delay, out-of-order delivery, etc.
30
UDP Ports:
multiplexing and demultiplexing
• Today’s operating systems support multi-processing, i.e.,
a number of applications can run simultaneously.
• Multiplexing:
— Each application program using UDP to send out data must first
negotiate with the operating system to obtain a UDP port.
— Any datagram the application program sends through the port
will have that port number in its UDP packet.
• Demultiplexing:
— UDP accepts incoming datagrams from the IP module, and
demultiplexes based on the UDP destination port.
31
UDP Applications
• UDP’s advantages
— Finer application-level control over what data is sent, and when
— No connection establishment: saves time!
— No connection state
— Small packet header overhead: only 8 bytes
— Can support multicast
• Disadvantage: UDP is unreliable
• Applications using UDP:
— Streaming multimedia
— Internet Telephony
— Network management
— Routing protocols
— DNS
32
TCP
•
RFC 793
•
TCP provides reliable communication between pairs of processes (by using
positive acknowledgement with retransmission).
•
— http://www.ietf.org/rfc/rfc793.txt
— No data loss, no error, no out-of-order (from the application’s point of view).
— TCP uses the idea of retransmission to recover packet loss.
TCP is connection-oriented stream service.
— From the hosts’ point of view, two applications need to “setup” a TCP
connection (like virtual circuit) before they use TCP to exchange data.
• But from the network’s point of view, they are still processing normal IP packets.
Routers don’t distinguish between UDP packets and TCP packets.
— TCP connection is full-duplex.
— A TCP connection is recognized by (IPs, IPD, Ports, PortD, “TCP” )
— TCP only supports point-to-point communication between TWO hosts. It cannot
support multicast or broadcast which can be supported by UDP.
— Data is regarded as byte stream. The TCP must deliver exactly the same
sequence of bytes to the receiver application, though the IP packets may arrive
out of order.
— The stream is unstructured. There is no record boundaries.
33
TCP/IP Concepts
34
TCP segment
• A TCP application generates data as a byte
stream.
• TCP module receives some bytes from the byte
stream, then composes a TCP segment by
adding a TCP header, then sends the segment
to IP module.
• The size of TCP header is at least 20 octets.
—TCP header can have some options.
35
TCP
• TCP is the most complicated protocol in
Internet. It is evolving all the time.
• More than 90% of today’s Internet traffic are
TCP. It has very wide applications.
• You need to take another course to learn the
details of TCP/IP:
—COMP3040: Internet & the World Wide Web
36
Application Layer
• Contains the logic needed to support the user applications
• Usually, each type of application needs to have one or more
protocols.
• E.g.
— Email applications
• Sending email -- smtp: [http://www.ietf.org/rfc/rfc821.txt]
• Accessing email -- pop3: [http://www.ietf.org/rfc/rfc1939.txt]
— telnet
• [http://www.ietf.org/rfc/rfc854.txt]
— File transfer
• FTP: File transfer protocol, [http://www.ietf.org/rfc/rfc959.txt]
— Web application
• HTTP/1.0: [http://www.ietf.org/rfc/rfc1945.txt]
• HTTP/1.1: [http://www.ietf.org/rfc/rfc2616.txt]
— P2P file sharing
• Lots of different protocols, like BitTorrent, eDonkey
• You can design and implement your own!
37
Some Protocols in TCP/IP Suite
38
Addressing Issue
• “Address” is used to identify an object
— It is common to use several addresses together to identify an object
• Process level address: to identify a process
— Port number (TCP/UDP)
• Network level address: to identify a host
— IP address (IP)
— Unique IP address for each end system (computer) and router. A router
has more than one IP addresses, each for a different interface.
• Link level address: to identify a network card
— Physical address (MAC address)
• On Windows XP DOS prompt, type:
— C:\>ipconfig /all
— C:\>netstat -a
39
Trace of Simple Operation
• Process associated with port 1 in host A wants to send
message to port 2 in host B
— Process at A hands down message to TCP module, with
instructions to send it to host B, port 2
— TCP appends a TCP header, hands down to IP module to send to
host B, with instructions to send it to host B
— IP appends an IP header, hands down to network access layer
(e.g. Ethernet) to send to router J, with instructions to send it
to router J
— Network access layer appends link header, sends out the bits to
the connected router J through the network interface
— At router J, the link header is stripped off and the IP header
examined. From the IP header, router J knows that this packet
is destined to host B, and actions accordingly.
40
Protocol Data Units (PDUs)
in TCP/IP Architecture
41
Section 3
The OSI Model
42
OSI Reference Model
• OSI: Open Systems Interconnection
• It was developed by the International
Organization for Standardization (ISO), starting
from 1977.
• Seven layers later, OSI was published as ISO
standard, ISO 7498, in 1984.
• A theoretical system delivered too late!
• TCP/IP has become the de facto standard for
data communications.
43
OSI - The Model
• A layer model
— Each layer performs a related subset of the functions required to
communicate with another system.
— Each layer relies on the next lower layer to perform more
primitive functions and to conceal the details of those functions.
— Each layer provides services to the next higher layer.
— Ideally, changes in one layer do not require changes in other
layers.
• The task of ISO was to define a set of layers and the
services performed by each layer.
— The partitioning should group functions logically and should
have enough layers to make each layer manageably small, but
should not have so many layers that the processing overhead is
burdensome.
44
OSI Layers
Seven layers have been
defined in OSI architecture.
45
The OSI Environment
46
OSI Layers (1)
• Physical Layer
—Physical interface between devices
•
•
•
•
Mechanical
Electrical
Functional
Procedural
• Data Link Layer
—Means of activating, maintaining and deactivating a
reliable link
—Error detection and control
—Higher layers may assume error free transmission
47
OSI Layers (2)
• Network Layer
— Transport of information
— Higher layers do not need to know about underlying technology
— Not needed on direct links
• Transport Layer
— Exchange of data between end systems
— Error free
— In sequence
— No losses
— No duplicates
— Quality of service
48
Network layer: use of a Relay
49
OSI Layers (3)
• Session Layer
—Control of dialogues between applications
—Dialogue discipline
—Grouping
—Recovery
• Presentation Layer
—Data formats and coding
—Data compression
—Encryption
• Application Layer
—Means for applications to access OSI environment
50
OSI v TCP/IP
51
KEY POINTS
• A protocol architecture is the layered structure of
hardware and software that supports the exchange of
data between systems and supports distributed
applications, such as electronic mail and file transfer.
• At each layer of a protocol architecture, one or more
common protocols are implemented in communicating
systems. Each protocol provides a set of rules for the
exchange of data between systems.
• The most widely used protocol architecture is the
TCP/IP protocol suite, which consists of the following
layers: physical, network access, internet, transport, and
application.
• Another important protocol architecture is the sevenlayer Open Systems Interconnection (OSI) model.
52