This Lecture BUS3150 - Computer Facilities Network ......2 - Network Models - OSI and TCP/IP 17 Faculty of Information Technology Operation of a Protocol Architecture The next step

� �

� �

2 - Network Models - OSI and TCP/IP

Faculty of Information Technology

BUS3150 - Computer Facilities Network Management


Faculty of Information Technology

Monash University

� �

� �


2 Faculty of Information Technology

This Lecture

• Protocols and protocol architecture fundamentals.

– An example three layer model.

– Addressing requirements.

– Protocol data units (PDU).

• The OSI Model.

• The TCP/IP Model.

� �

� �



Protocols

• Used for communications between entities in a system.

• Must speak the same language.

• Entities:

– Anything capable of sending/receiving information.

– Examples: User applications, DBMS, e-mail etc.

• Systems:

– A physically distinct object that contains entities.

– Examples: Terminal, computers, remote sensors.

• Protocols: a set of rules governing exchange of data between entities.

� �

� �



Need For Protocol Architecture

• Consider a file transfer between two computers:

1. Source must activate a direct communications path or inform the net-work of the destination.

2. Source must check the destination is prepared to receive data.

3. File transfer application on source must check destination file manage-ment system will accept and store file for his user

4. May need file format translation.

• Instead of implementing above logic in a single module, it is broken upinto a vertical stack of sub tasks.

• Each layer in the stack performs a related subset of the function requiredto communicate with the other system.

� �

� �



Need For Protocol Architecture

• Ideally the layers should be defined so that changes in one layer do notrequire changes in another layer.

– A lower layer performs more primitive functions and provides a serviceto the higher layer.

• It is obvious that for two systems to communicate, the layered functionsmust exist in both systems.

– That is, communication is achieved by having the peer, or correspond-ing layers in the two systems to communicate.

• For example file transfer could use three layers:

– File transfer application (tasks 3 and 4).

– Communication service module (task 2).

– Network access module (task 1).

� �

� �



Simplified File Transfer Architecture

Networkinterface logic

Networkinterface logic

Communications−related messages

Files and file transfer commands

Computer YComputer X

Network accessmodule

Communicationsservice module

File transferapplication

Network accessmodule

Communicationsservice module

File transferapplication

C

C

o

o

m

m

m

m

u

u

n

n

i

i

c

c

a

a

t

t

i

i

o

o

n

n

s

s

n

n

e

e

t

t

w

w

o

o

r

r

k

k

� �

� �



Key Elements of a Protocol

• Peer layer communication takes place using formatted blocks of data thatobey a set of rules or conventions known as a protocol.

• Syntax:

– Data formats: what fields do the blocks of data contain?

– Signal levels.

• Semantics

– Control information: acknowledgements from the receiver that the filetransfered without corruption.

– Error handling: if not received correctly, then retransmit the file.

• Timing

– Speed matching.

– Sequencing.

� �

� �



A Three Layer Model

• Generally, communications can be said to involve three agents: Applica-tions, Computers and Networks.

– Applications execute on computers that typically support multiple si-multaneous applications.

– Computers are connected to networks.

– Data to be exchanged are transferred by the network.

• With these concepts in mind, it appears natural to organise communicationtask into three relatively independent layers:

– Network access layer.

– Transport layer.

– Application layer.

� �

� �



Protocol Communication

Networkaccess protocol

Networkaccess protocol

Transport Protocol

Application Protocol

Computer Y

Protocols in a Simplified Three Layer Architecture

Network access

Transport

Application

Computer X

Network access

Transport

Application

Communicationsnetwork

� �

� �



Network Access Layer

• Network access layer is concerned with the exchange of data between acomputer and the network to which it is attached.

• The software in this layer depends on the type of network used such ascircuit switched, packet switched, LANs etc.

• By separating this function into a separate layer, the remainder of commu-nication software, above this layer, need not be concerned about specificsof the network used.

• Each computer at this layer requires a unique network address.

� �

� �



Transport Layer

• Regardless of the application, there is a need that data is exchanged reli-ably.

• That is, all of the data should arrive at the destination and in the sameorder in which it was sent.

• As the need for reliability is independent of the application, it makes senseto collect those mechanisms in a common layer shared by all applications.

• This common layer is referred to as the transport layer.

� �

� �



Application Layer

• Application layer contains the logic needed to support various applica-tions.

– For different applications, a separate module is needed.

– For example, one for file transfer, another for email, etc.

• Each application on a computer must have an address that is unique withinthat computer.

– This allows the transport layer to support multiple application on a sin-gle computer.

– Each application that accesses the services of the transport layer is iden-tified as a service access point (SAP) or port.

� �

� �



Addressing Requirements

• Two layers of addressing: network addresses and application addresses.

Network access

Computer A

Computer C

Computer B

Transport

Applications1 2 3 4

Service access point

Network address

( ) ( ) ( ) ( )

Network access

Transport

Applications1 2 3

( ) ( ) ( )

Network access

Transport

1 2Applications

( ) ( )

Communicationsnetwork

� �

� �



Protocol Communication

• Suppose that an application on computer A (associated with SAP3) needto send a message to an application (associated with SAP1) on ComputerB.

• The application on A hands the message over to the transport layer withinstructions to send it to SAP1 on computer B.

• The transport layer hands the message over to the network access layer,which instructs the network to send the message to computer B.

– The network need not be told the identity of the destination SAP.

– All that it needs to know is that data is intended for computer B.

• To control the above operation, control information, as well as user datamust be transmitted as shown in the next slide.

� �

� �



Operation of a Protocol Architecture

Application

Transport

NetworkAccess

Source X

Application

Transport

NetworkAccess

Destination Y

DHost

DSAP

DSAP = destination service access pointDHost = destination host

Record

DHost

DSAP

Record

Transport PDU

Packet

� �

� �




• The sending application generates a block of data and passes it to thetransport layer.

• The transport layer appends a transport header containing protocol controlinformation.

• The combination of data from the previous layer and control informationis known as protocol data unit (PDU).

• In this example it is known as transport PDU.

• The header in a transport PDU contains control information to be used bythe peer transport protocol at computer Y. For example:

– Destination SAP

– Sequence number

– Error-detection code

� �

� �




• The next step is for the transport layer to hand each PDU over to thenetwork layer with instruction to transmit it to the destination computer.

• The network access protocol appends a network access header to the datait receives from transport layer, creating network access PDU.

• Examples of the items that may be stored in the header include:

– Destination computer address

– Facilities requestApplication Data

TransportHeader

TransportHeader

NetworkHeader

Transportprotocol data units

Networkprotocol data units

(packets)

NetworkHeader

� �

� �



Standardised Protocol Architectures

• When communication is desired among computers from different vendors,the software development effort can be a nightmare.

– Different vendors use different data formats and data exchange proto-cols.

– Even within one vendor, different model computers may communicatein unique ways.

• As the use of computer communications and computer networking prolif-erates, vendors are forced to adopt a common set of conventions.

� �

� �



Standardised Protocol Architectures

• Two protocol architectures have served as the basis for the developmentof interoperable protocol standards:

– OSI reference model

· Well known but has never lived up to its early promise.

– TCP/IP protocol suite

· By far the most widely used interoperable architecture.

� �

� �



OSI Reference Model

• A widely accepted structuring technique, layering, is used by ISO forOpen Systems Interconnection (OSI) model.

– The communications functions are partitioned into a hierarchical set oflayers.

– Each layer relies on the next lower layer to perform more primitivefunctions.

– Each layer provides services to the next higher layer.

– Ideally, the layers should be defined so that changes in one layer do notrequire changes in other layers.

· The problem is divided in to manageable sub-problems.

� �

� �



OSI Reference Model

• The task of ISO in defining OSI was to define a set of layers and theservices performed by each of them.

• The partitioning should group functions logically and should have enoughlayers to make each layer manageably small.

• Should not have too many layers so that processing overhead is burden-some.

• The resulting reference model has 7 layers as shown in the next slide.

� �

� �



OSI Layers

Provides independence to the application processes from differences in datarepresentation (syntax).

Presentation

Provides control structure for communication between applications; establishes,manages and terminates connections (sessions) between cooperating applications.

Session

Provides reliable, transparent transfer of data between end points; providesend−to−end error recovery and flow control as well.

Transport

Provides upper layers with independence from data transmission and switchingtechnologies used to connect systems.

Network

Provides reliable transfer of information across the physical link; sends blocks(frames) with the necessary synchronisation, error control and flow control.

Data Link

Concerned with transmission of unstructured bit stream over physical medium;mechanical, electrical, functional, and procedural characteristics of medium.

Physical

Provides access to the OSI environment for users and also provides distributedinformation services.

Application

� �

� �



OSI Architecture

• Each of the two systems that communicate contains the seven layers.

• An application on one machine invokes the application layer in the other.

– The two layers on the two machines establish a peer relationship.

• These peer relationships are created in all layers down to the physicallayer.

• However, there is no direct communication between peer layers except atthe physical layer.

� �

� �



OSI Architecture

• Even at the physical layer, the OSI model does not stipulate that the twosystems be directly connected.

– A packet switched or circuit switched network may be used to providethe communication link.

• Similar to the three layer model discussed earlier, Protocol Data Units(PDUs) are created in all bottom 6 layers encapsulating user data.

• At each layer, data units may be segmented into several parts to accom-modate its requirements.

� �

� �



The OSI Environment

Application

Presentation

Session

Transport

Network

Data Link

Physical

Application

Presentation

Session

Transport

Network

Data Link

PhysicalDL−PDU

N−PDULH LT

NH T−PDU

TH S−PDU

SH P−PDU

PH A−PDU

AH User Data

DL−PDU

N−PDULH LT

NH T−PDU

TH S−PDU

SH P−PDU

PH A−PDU

AH User Data

CommunicationsPath (e.g., point−

to−point link,network)

OutgoingPDU Construction

IncomingPDU Reduction

Application X Application Y

� �

� �



Standardisation within OSI Framework

• The principal motivation for development of the OSI model was to providea framework for standardisation.

• The model facilitates the standards-making process in two ways:

– As the functions of each layer are well defined, standards can be de-fined independently and simultaneously.

– As the boundaries between layers are well defined, changes in stan-dards in one layer need not affect already existing software in anotherlayer.

� �

� �




• The principles used essentially amount to using modular design.

• The lower layers are concerned with greater detail while the upper layersare independent of these details.

• Each layer provides services to the next higher layer and implements aprotocol to the peer layer in other systems.

� �

� �




Layer 7(Application)

Layer N

Layer 1(Physical)

TotalCommunication

Function Decompose(modularity,

information−hiding)

Layer Nentity

Service toLayer N+1

Service fromLayer N−1

Protocolwith peerLayer N

OSI−wide standards(e.g., network management, security)

The OSI Architecture as a Framework for Standardization

Addressing(Service Access Point)

Protocol Specification(Precise syntax and

semantics forinteroperability)

Layer N

Service Definition(Functional description

for internal use)

Layer−Specific Standards

� �

� �




• The nature of the standardisation required at each layer can be describedusing three key elements.

• Protocol specification:– Two entities at the same layer in different systems cooperate and inter-

act by means of a protocol.– As two different open systems are involved, the protocol must be spec-

ified precisely.• Service definition:

– In addition to the protocols that operate at a given layer, standards areneeded for services that each layer provides to the next higher layer.

– Typically, the definition of services defines what services are provided,but not how they are provided.

• Addressing:– Each layer provides services to entities at the next higher layer by

means of service access points (SAPs).

� �

� �




• The interaction between two adjacent layers takes place within the con-fines of a single system and is not the concern of any other open system.

• Thus, as long as peer layers in different systems provide the same serviceto their next higher layers, the details of how the services are providedmay differ without loss of interoperability.

� �

� �



Physical Layer

• Concerned with transmission of an unstructured bit stream over a physicalmedium from one hop (node) to the next.

• Deals with the mechanical, electrical, functional, and procedural charac-teristics to access the physical medium.

� �

� �



Data Link Layer

• Provides for the reliable transfer of information across the physical linkincluding activation, maintenance and deactivation of the link.

• Sends blocks (frames) with the necessary synchronisation, error control,and flow control - Higher layers may assume error free transmission.

� �

� �



Data Link Layer Example� �

� �



Network Layer

• Responsible for end-to-end delivery of a data packet.

• Provides upper layers with independence from the data transmission, switch-ing and routing technologies used to connect systems.

• Establish, maintain, and terminate connections across a network (i.e. throughnodes).

• Has view of the entire network (data link and physical layers address sub-networks only).

� �

� �



Network Layer Example� �

� �



Use of a Relay

7

6

5

4

3

2

1

7

6

5

4

3

2

1

Physical media for OSI Physical media for OSI

Application−layer protocol

Presentation−layer protocol

Session−layer protocol

Transport−layer protocol

Network

Data Link

Physical

Network

Data Link

Physical

ENDSYSTEM

RELAYSYSTEM

ENDSYSTEM

� �

� �



Transport Layer

• Provides reliable, transparent transfer of data between processes on theend points - error free, in sequence, no losses, no duplicates.

• Provides end-to-end error recovery and flow control.

• May also be concerned with optimising network use and quality of ser-vice.

� �

� �



Transport Layer Example

� �

� �



Session Layer

• Provides the control structure for communication between applications in end systems.

• Establishes, manages, and terminates connections (sessions) between cooperating appli-

cations.

• May provide services such as interruption and recovery of dialog through the use of

checkpoints.

� �

� �



Presentation Layer

• Provides independence to the application processes from differences indata representation (syntax).

• Data formats and coding, data compression, encryption.

� �

� �



Application Layer

• Provides access to the OSI environment for users and also provides dis-tributed information services.

� �

� �



TCP/IP Protocol Architecture

• The TCP/IP protocol architecture is a result of protocol research and de-velopment on the experimental packet-switched network, ARPANET

– It was funded by Defence Advanced Projects Agency (DARPA) andgenerally referred to as the TCP/IP protocol suite

• This protocol suite consists of a large collection of protocols that havebeen issued as Internet standards by the Internet Architecture Board (IAB)

� �

� �



TCP/IP Layers

• The TCP/IP model organises the communication task into five relativelyindependent layers:

Physical

NetworkAccess

Internet

Application

Transport(host−to−host)

TCP/IP

Application

Presentation

Session

Transport

Network

Data Link

Physical

OSI

� �

� �



TCP/IP Layers

• The physical layer covers the physical interface between a data transmis-sion device and a transmission medium or network.

– Specifies the characteristics of the transmission medium, the nature ofthe signals, the data rate etc.

• The network access layer is concerned with the exchange of data betweenan end system and the network to which it is attached.

– The software used at this layer depends on the type of network to beused (circuit switching, packet switching, LANs etc).

� �

� �



TCP/IP Layers

• The network access layer is concerned with routing data across a networkfor two end systems attached to the same network.

• In the cases where two devices are attached to different networks, proce-dures are required to allow data to traverse multiple interconnected net-works.

– This is the function of the Internet layer, which uses the Internet Proto-col (IP) for it.

– This protocol is used not only in the end systems, but also in routers (aprocessor that connects two networks).

� �

� �



TCP/IP Layers

• The transport, or host-to-host layer is responsible for reliable exchange ofdata between end systems, regardless of the nature of the application.

– The transmission Control Protocol (TCP) is the most commonly usedprotocol for this functionality.

• The application layer contains the logic needed to support various userapplications.

� �

� �



TCP and UDP

• For most applications running as part of the TCP/IP architecture, the trans-port layer protocol is TCP.

• TCP provides a reliable connection in the form of a temporary logicalassociation to transfer data between applications.

• A TCP PDU is called a TCP segment, which includes source port anddestination port in its header.

– Ports serve the same service as SAPs in the OSI model, and identifiesthe respective user applications.

� �

� �



TCP and UDP

• In addition to TCP, the other commonly used transport layer protocol thatcomes as part of TCP/IP suite is the User Datagram Protocol (UDP).

• UDP does not guarantee delivery, preservation of sequence or protectionagainst duplication.

• UDP enables a procedure to send messages to other procedures with min-imum of protocol mechanism / overhead.

• Essentially UDP adds port addressing capability to IP.

� �

� �



Operation of TCP and IP

Router J

TCP

IP

Physical Physical

IP

NAP 1 NAP 2

Physical Physical

Network AccessProtocol #1

Host A

App XApp Y

TCP

IP

Network AccessProtocol #2

Host B

App YApp X

Network 1 Network 2

Global networkaddress

Subnetwork attachmentpoint address

Logical connection(e.g., virtual circuit)

Logical connection(TCP connection)

Port orservice access point (SAP)

� �

� �




• Communication facility that uses TCP/IP may consist of multiple net-works, which are referred to as subnetworks.

• Some sort of a network access protocol, such as Ethernet, is used to con-nect a computer to a subnetwork.

– This protocol enables a host to send data across the subnetwork to an-other host or, if the target host is on another subnetwork, to a routerthat will forward the data.

• IP is implemented in all of the end systems and routers while TCP isimplemented only in the end systems.

� �

� �




• For successful communication, every entity in the overall system musthave a unique address.

• Two levels of addressing are needed for this:

– Each host on a subnetwork must have a unique global internet address.

· This allows the data to be delivered to the proper host.

– Each process with a host must have an address that is unique within thehost.

· This allows host-to-host protocol (TCP) to deliver data to the properprocess.

� �

� �




User data

TCPheader

IPheader

Networkheader

Applicationbyte stream

TCPsegment

IPdatagram

Network−levelpacket

� �

� �




• As shown in the previous slide, to accomplish data transmission, controlinformation as well as user data must be transmitted.

• The control information is added at each layer by means of a header ap-pended.

• Examples of items in a TCP header includes...

• Destination port:

– When the TCP entity at the receiver receives data, it must know wherethe data to be delivered.

� �

� �




• Sequence number:

– TCP numbers the segments it sends it sends to a particular destinationport sequentially.

– As a result if the segments arrive at the receiver out of order, they canbe reordered.

• Checksum:

– The sending TCP includes a code that is a function of the contents ofthe remainder of the segment.

– The receiver TCP performs the same calculation, and compares theincoming code for any error detection.

� �

� �




• An example of an item stored in the IP header is the destination host ad-dress.

• The header appended at the network access layer contains the informationthat is needed to transfer data across the subnetwork.

• Examples of the items that may be contained in this header includes:

– Destination address within the subnetwork.

– Facilities requests such as priority.

� �

� �




• At a router, the packet header (network header) is stripped off and the IPheader is examined.

• On the basis of the destination address in the IP header, the datagram isdirected out across the subnetwork.

– To achieve this, the datagram is again augmented with a network accessheader.

• Some applications that have been standardised to operate on top of TCPare Simple Mail Transfer Protocol (SMTP), File Transfer Protocol (FTP)and TELNET.

� �

� �



Physical Address� �

� �



IP Address

� �

� �



Port Address� �

� �



Some Protocols in the TCP/IP Suite

BGP = Border Gateway ProtocolFTP = File Transfer ProtocolHTTP = Hypertext Transfer ProtocolICMP = Internet Control Message ProtocolIGMP = Internet Group Management ProtocolIP = Internet ProtocolMIME = Multi−Purpose Internet Mail Extension

OSPF = Open Shortest Path FirstRSVP = Resource ReSerVation ProtocolSMTP = Simple Mail Transfer ProtocolSNMP = Simple Network Management ProtocolTCP = Transmission Control ProtocolUDP = User Datagram Protocol

FTP HTTP

TCP

SMTP

MIME

TELNET SNMP

UDP

IGMP OSPF

BGP

IP

RSVPICMP

Documents

This Lecture BUS3150 - Computer Facilities Network ......2 - Network Models - OSI and TCP/IP 17 Faculty of Information Technology Operation of a Protocol Architecture The next step