Data
`
`Communications
`and Networks
`3rd Edition
`
`Edited by R.L. Brewster
`
`The Institution of Engineering and Technology
`
`Petitioners’ EX. 1017 — Page 1
`
`

`
`Published by The Institution of Engineering and Technology, London, United Kingdom
`
`First edition © 1994 The Institution of Electrical Engineers
`Reprint with new cover © 2008 The Institution of Engineering and Technology
`
`First published 1994
`Reprinted 2008
`
`This publication is copyright under the Berne Convention and the Universal Copyright
`Convention. All rights reserved. Apart from any fair dealing for the purposes of research
`or private study, or criticism or review, as permitted under the Copyright, Designs and
`Patents Act, 1988, this publication may be reproduced, stored or transmitted, in any
`form or by any means, only with the prior permission in writing of the publishers, or in
`the case of reprographic reproduction in accordance with the terms of licences issued
`by the Copyright Licensing Agency. Inquiries concerning reproduction outside those
`terms should be sent to the publishers at the undermentioned address:
`
`The Institution of Engineering and Technology
`Michael Faraday House
`Six Hills Way, Stevenage
`Herts, SGi 2AY, United Kingdom
`
`\Nvvw.theiet.org
`
`While the author and the publishers believe that the information and guidance given
`in this work are correct, all parties must rely upon their own skill and judgement when
`making use of them. Neither the author nor the publishers assume any liability to
`anyone for any loss or damage caused by any error or omission in the work, whether
`such error or omission is the result of negligence or any other cause. Any and all such
`liability is disclaimed.
`
`The moral rights of the author to be identified as author of this work have been
`asserted by him in accordance with the Copyright, Designs and Patents Act 1988.
`
`British Library Cataloguing in Publication Data
`A catalogue record for this product is available from the British Library
`
`ISBN (10 digit) 0 85296 304 3
`ISBN (13 digit) 918-0-85296-804-8
`
`Printed in the UK by Redwood Books, Trowbridge, Wiltshire
`Reprinted in the UK by Lightning Source UK Ltd, Milton Keynes
`
`Petitioners‘ Ex. 1017 - Page 2
`
`

`
`The philosophy of the OSI seven-layer model
`
`2 7
`
`The aim of the ISO Reference Model is to provide a framework for the
`coordination of standards development and to allow existing and evolving standards
`activities to be set within it common framework. The aim is to allow an application
`process in any com uter that supports a particular set of standards to communicate
`freely with an app ication process in any other computer that supports the same
`standards, irrespective of its origin of manufacture.
`Some examples of application processes that may wish to communicate in an open
`way are:
`
`0
`
`0
`
`0
`0
`0
`
`0
`
`0
`
`a process (program) executing in a computer and accessing a remote file
`system;
`a process acting as a central file service (sewer) to a distributed community of
`(client) processes;
`‘
`a prpcess in an office workstation (computer) accessing an electronic mail
`SCIVCC;
`a process acting as an electronic mail server to a distributed community of
`(c ient) processes;
`a process in a supervisory computer controlling a distributed community of
`computer-basedinstruments or robot controllers associated with it process or
`automated manufacturing plant;
`a process in an instrument or robot controller receiving commands and
`returning results to a supervisory system;
`a process in a bank computer that initiates debit and credit operations on a
`remote system.
`
`Open systems interconnection ‘is concerned with the exchange of-information
`between such processes.
`'l:he'aim is to enable‘ application processes to cooperate in
`carrying out a particular (distributed) information processing task irrespective of the
`computers on which they are running.
`
`3.3 ISO Reference Model
`
`A communication subsystem is a complex piece of hardware and software. Early
`attempts at implementing the software for such subsystems were often based on a
`single, complex, unstructured program (normally written in assembly language) with
`many interacting components. The resulting software was difficult to test and often
`very difficult to modify.
`To overcome this problem, the ISO has adopted a layered approach for the
`reference model. The complete communication subsystem is broken down into a
`number of layers each of which performs a well defined function. Conceptually.- these
`layers can be considered as performing one of two generic functions; network-
`dependent functions and application-oriented functions. This in turn gives rise to three
`distinct operational environments:
`
`1. The network environment, which is concerned with the protocols and
`standards relating to the different types of underlying data communication
`networks.
`
`2. The OSI environment, which embraces the network environment and adds
`additional application-oriented protocols and standards to allow end systems
`(computers) to communicate with one another in an open way.
`
`Petitioners‘ Ex. 1017 - Page 3
`
`

`
`2 8
`
`The philosophy of the OSI seven-layer model
`
`3. The real systems environment, which builds on the OSI environment and is
`concerned with a manufacturer's own proprietary software and services which
`have been developed to perform a particular distributed information processing
`task.
`
`This is shown in diagrammatic form in Fig. 3.2.
`
`Application
`oriented
`functions
`
`Network
`
`dgpendem
`functions
`
`Application
`
`functions
`
`NelW0l'|<
`
`dependent
`functions
`
`Data network
`
`Network environment
`
`OSI environment
`
`Real s stems environment
`
`Fig. 3.2 Operational environments
`
`Both the network-dependent and application-oriented (network-independent)
`components of the OSI model are implemented as a number of layers. The boundaries
`between each layer, and the functions performed by each layer, have been selected on
`the basis of experience gained during earlier standardisation activity.
`Each layer performs a well defined function in the context of the overall
`communication subsystem. It operates according to a defined protocol by exchangittg
`messages. both user data and additional control information. with a corresponding
`peer layer in a remote system. Each layer has a well defined interface between itself
`and the layer immediately above and below. Consequently. the implementation of a
`particular protocol layer is independent of all other la ers.
`The logical structure of the ISO Reference M el is made up of seven protocol
`layers as shown in Fig. 3.3.
`The three lowest layers (1-3) are network-dependent and are concerned with the
`protocols associated with the data communication network being used to link the two
`communicating computers. In contrast, the three upper layers (5-7) are application-
`oriented and are concerned with the protocols that allow two end user application
`processes to interact with each other, normally through a range of services 0 feted by
`the local operating system. The intermediate transport layer (4) masks the upper
`application-oriented layers from the detailed operation of the lower network-dependent
`layers. Essentially, it builds on the services provided by the latter to provide the
`application-oriented layers with a network-independent message interchange service.
`The function of each layer is specified formally as a protocol that defines the set of
`rules and conventions used by the layer to communicate with a similar peer layer in
`another (remote) system. Each layer provides a defined set of services to the layer
`immediately above. It also uses the services provided by the layer immediately below
`
`Petitioners‘ Ex. 1017 - Page 4
`
`

`
`The philosophy of the OSI seven-layer model
`
`29
`
`A-L<7>
`—
`<---—--——> P-we
`+-----—+ S-L<5>
`T-M4)
`N—L<3>
`L-L(2)
`
`Link Layer
`
`Ph sical La er
`
`P-L l
`
`Network environment
`
`OSI environment
`
`Real systems environment
`
`Fig 3.3 Overall structure ofthe ISO Reference Model
`
`it to transport the message units associated with the protocol to the remote peer layer.
`For example. the transport layer provides a network-independent message transport
`service to the session layer above it and uses the service provided by the network layer
`below it to transfer the set of message units associated with the trans
`rt protocol to a
`peer transport‘ layer in another system. Conceptually.
`there are, each layer
`communicates with a similar peer layer in a remote system according to a defined
`protocol. However, in practice the resulting protocol message units of the layer are
`passed by means of the services provided by the next lower layer. The basic functions
`of each layer are summarised in Fig. 3.4.
`
`3.3.1 The application-oriented layers
`
`3.3.1 .I The application layer
`The application layer provides the user interface - normally an application
`program/process - to a range of network-wide distributed information services. These
`include fi e transfer access and management as well as general document and message
`interchange services such as electronic mail. A number of standard protocols are either
`available or are being developed for these and other ty‘ es of service.
`Access to application services is normally ach eved through a defined set of
`primitives, each with associated parameters. which are supported by the local
`operating system. The access primitives are the same as other operating system calls
`(as used for access to, say, a local file system) and result in an appropriate operating
`system procedure (process) being activated. These operating system procedures use
`
`Petitioners‘ Ex. 1017 - Page 5
`
`

`
`LANS: Protocols above the medium access layer
`
`1 09
`
`The Control field is 16 bits long for frames containing sequence numbers (data and
`flow control (supervisory) frames for the connection-oriented.service), offering 7-bit
`sequence numbers. Un-numbered frames have 8-bit control fields, The full list of
`frame types is given in Table 8.2. For the connectionless service, three frame types
`have been introduced: UI for connectionless data, XID which allows LLC
`implementations to exchange information on their capabilities. TEST which tests that
`communication exists between LLC implementations and two acknowledgement frame
`types for the acknowledged connection ess service.
`
`8.3 The Internet protocols
`
`The Internet protocol suite, supported by the USA Department of Defense (DoD). has
`a long history. The Defense Advanced Projects Research Agency (DARPA) Internet is
`a large collection of interconnected networks of many types including various LANs
`and WANs and terrestrial, satellite and radio links. It has evolved from a number of
`separate networks, principally the Arpanet, a research network which grew rapidly
`during the 1970s and 19808. The Internet connects academic, military and other
`organisations and now extends beyond the USA to include countries such as Australia
`and is directly connected to many other networks internationally.
`One reason for the importance of the protocol suite is that it must be supported by
`any equipment attached to the Internet. This has put pressure on computer suppliers to
`offer the protocols. which are now available for a large range of computers from
`personal computers and workstations to mainframes. Recently, the UK JANET
`network (Joint Academic NETwork) has added the Internet suite to its range of
`protocols and it is now possible to communicate directly from systems attached to
`JANET to machines on the Internet.
`
`OS! layer
`
`DARPA Internet protocols
`
`Application
`
`Presentation
`
`Network
`
`Data Link
`
`-
`
`F'l°(g;.‘;,n)sfcr
`
`Terminal
`
`emulation
`(TELNET)
`
`.
`
`El°§;‘:"‘°
`
`Other
`
`application
`protocols
`
`Transmission Control
`Protocol (TCP)
`
`User Datagram
`Protocol (UDP)
`
`Internet Protocol
`
`_
`
`Address Resolution
`PTOIOCOI
`
`I
`
`Miégfgeé gglgfgl
`
`Physical
`
`Local Area Network
`
`Fig. 8.3 The Internet protocol suite
`
`Petitioners’ Ex. 1017 - Page 6
`
`

`
`I I 0
`
`LANS: Protocols above the medium access layer
`
`The protocol suite is designed to ease communication between different types of
`networks and is based on the datagram (connectionless) approach. The various
`elements of the protocol suite are shown in Fig. 8.3, which also shows the rough
`correspondence with the OSI reference model. (Note that the figure shows examples
`of facilities at each layer but does not imply direct vertical relationships between
`individual components in the Internet suite.)
`The details of the Internet or TCP/IP suite and a number of associated documents
`are published by the DoD and are freely available across the Internet. They arepart of a
`collection of Request for Comments (RFCs) which also cover many other networking
`topics. A list of the principal RFCs related to the TCP/IP protocol suite is included in
`the reference section at the end of this chapter.
`.
`The Internet suite runs on many different LANs. It is possible to mount IP directly
`on top of the MAC layer of a LAN, but the current recommendation is to use the IEEE
`802.2 LLC connectionless service where possible, using the Un-numbered
`Information frames and s ecified values for DSAP and SSAP addresses etc. (The
`details are given in RFC 1 42.)
`At the network layer, the Internet Protocol (IP) enables individual datagrams to be
`sent across the network. It offers a number of addressing options and message
`fragmentation and reassembly. Also at this layer are the Address Resolution Protocol
`(ARP) [RFC 826] whose job is to map Internet addresses to Ethernet (CSMAICD)
`addresses. ARP uses broadcast transmission to advertise the system’s own address
`and to request information from other systems on the network. The Internet Control
`Message Protocol (ICMP) [ RFC 792] enables IP entities to exchange control and
`maintenance information.
`The transport layer consists of two sets of services accessed directly to the
`application. The Transmission Control Protocol (TCP) offers bidirectional reliable byte
`streams and employs error detection and correction and flow control procedures. The
`User Datagram Protocol (UDP) offers a connectionlessrservice at this layer.
`A number of standard applications are offered which use either TCP or UDP.
`These include file transfer, electronic mail etc. We shall now look at some of the
`elements of the Internet protocol suite in more detail.
`
`8.3.1 The Internet Protocol (IP)
`
`[RFC 791]
`
`The IP protocol is best described by reference to the IP header (Fig. 8.4) which is
`prepended to each data unit to be sent across the network (Fig. 8.5 shows the position
`of the IP header related to the rest of the protocol hierarchy.)
`There arc length fields for both the header (in 32-bit words) and the complete
`frame, including the header (in octets). The type of service consists of flags to s ecify
`reliability, precedence, delay and throughput parameters. The identification fie d is a
`unique value for this datagram.
`A message which is too long to fit into 'a single data unit is fragmented and the
`offset indicates where this fragment starts in relation to the beginning of the message
`(in 64-bit units). The time to live file is decreased as an IP datagram travels across a
`network boundary; if the time to live has expired, the datagram is discarded. This
`prevents datagrams from circulating endlessly in the Internet. The protocol field
`identifies the protocol of the next highcr layer (TCP or UDP). The checksum offers
`early identification of header errors; it is recomputed at each gateway because the time
`to live field is updated. IP is an unreliable protocol so a checksum error leads to the
`datagram being discarded. The source and destination addresses will be discussed in
`the next paragraph. The options field is variable length and indicates required facilities
`such as a specified route to be taken. Padding ensures that the ll’ header falls on a 32-
`bit boundary. The data field is a multiple of 8 bits with a limit of 65,535 octets for the
`whole of the [P datagram (including the header).
`
`Petitioners‘ Ex. 1017 — Page 7
`
`

`
`1 1 4
`
`LANS: Protocols above the medium access layer
`
`..._____..... 32 bits _....__.___._..._.___.a...
`
`Data (variable number of octets)
`
`Fig. 8.8 UDP Protocol header
`
`8.3.5 Application layer protocols
`
`A number of standard application protocols are defined and are used widely on the
`Internet. The Telnet protocol allows users to use their local terminal to log on remotely
`to another machine. Telnet uses TCP and offers a network virtual terminal which
`enables the user’s real terminal to be described and thus understood by the host
`system. The FTP protocol enables suitably authenticated users to move through a
`remote file system, to examine the contents of directories and to send files to and
`receive files from the remote system. It uses two separate TCP connections: one for
`control messages and one for the data transfer. Electronic mail uses the Simple Mail
`Transfer Protocol (SMTP) which again makes use of the reliability offered by TCP. If
`the remote machine specified as the destination is not available. the message remains
`on the sending host until it can be transferred. In addition to these standard protocols.
`most implementations allow user code to be written at the application layer and to use
`either UDP or TCP to implement their own communication application.
`
`8.4 Bridging and routing
`
`It is frequently necessary to interconnect LANs and to connect LANs to Wide Area
`Networks (WANs) in order to gain access to other resources or to communicate with
`other eople. Fig. 8.9 shows some of the more common interconnection units which
`must
`e installed to achieve this interconnectivity. Sections of Local Area Networks
`which are connected at the physical layer are called repeaters; they simply regenerate
`and transmit the electrical signals on to the next network segment. Systems
`interconnecting LANs are generally called bridges and are connected at Layer 2; these
`systems need appropriate interfaces to both of the networks they connect. Bridges can
`be reasonably simple to implement as LANs share a common protocol at the LLC
`layer. Where LANs and WANs are interconnected, the mapping is done at layer 3 of
`the OSI model; these systems are known as gateways (the European term) or routers
`(the American term). They must take account of the many possible differences between
`the two different networks, which will include connectiomoriented or connectionless
`working, different maximum packet size, different addressing schemes etc. The term
`router reflects the fact that systems such as these must be able to determine where to
`send the information whether the destination system lies within one of the directly
`connected networks or on a network which is reached via one of them.
`
`Petitioners‘ Ex. 1017 — Page 8
`
`

`
`LANS: Protocols above the medium access layer
`
`1 15
`
` 1‘ Gateway
`
`
`or router
`
`Public WAN
`
`Fig. 8.9 Typical interconnection systems
`
`CSMAICD LAN
`
` Packet
`radio
`subnetwork
`
`
`
`I
`I
`|
`o
`,
`I
`
`IIIII
`
`‘
`g
`.
`'
`‘
`
`,
`
`,
`
`,,
`
`2
`
`.V
`
`._
`
`v
`
`
`L
`
`TCP header
`[P header
`CSMA/CD header
`
`=
`:
`:
`I
`'
`.
`'
`‘
`
`L
`TCP header
`IP header
`Pkt. radio hcadcr
`
`=
`:
`:
`|
`I
`|
`I
`C
`
`Fig. 8.10 Encapsulation of II’. datagrams at network routers
`
`Petitioners’ Ex. 1017 - Page 9
`
`

`
`I I 6
`
`LANS: Protocols above the medium access layer
`
`Use of the TCP/IP suite makes the functions of bridges and routers reasonably
`simple, once an appropriate route is known. The reliability achieved by TCP is end-to-
`end, so the routers need not be concerned with error recovery or flow control
`procedures. Fragmentation may be necessary when a datagram enters a network with a
`smaller maximum acket size than the one it left. In order to traverse a particular
`network. the IP pac et will be encapsulated in a protocol frame for each network as it
`enters the network and unwraPP'-d. by the router as it leaves the network. Both
`encapsulation and fragmentation are shown in Fig. 8.10.
`
`8.4.1 Bridges
`
`Bridges are reasonably simple; they forward frames as quickly as possible and provide
`buffering for where this is not immediately possible. If buffers overflow because
`congestion has occurred.
`frames may be discarded. Unfortunately. each IEEE 802
`LAN standard has a different frame format, so translation must be done by the bridge.
`They also have different data rates making congestion more likely when forwarding,
`for example, from a CSMA/CD network operating at 10 Mbit/s to a token ring
`operating at 4 Mbit/s. Their difference in frame lengths is serious as the MAC
`protocols do not offer segmentation and reassembly. so in some cases frames which
`are too long must be discarded.
`There are two princi al routing strategies for bridges [8]. The first is the
`transparent bridge, whic
`gradually learns the route to distant hosts. Initially all
`frames are flooded onto all possible connected networks. As each frame arrives its
`source address indicates where a particular host is situated, so that the bridge learns
`which way to forward future frames to that address. To prevent looping. the bridges
`form a s anning tree, connecting all bridges and offering a single route between
`bridges. n the source routing bridge. users must specify which route is to be
`followed. A route is found by sending a discovery frame which is replicated to use all
`possible paths to the destination; the frame collects bridge addresses as it passes
`through them. The destination responds to each discovery frame, enabling the source
`to chioosegthe most appropriate route. A comparison of the two techniques can be
`foun in [ ].
`
`8.5 Personal Computer networks
`
`in addition to the standard network technologies and protocols described in the IEEE
`802 standards. there are a number of proprietary approaches to LAN interconnectivity.
`Examples include:
`
`NETBIOS
`Novell’s Netware
`3-Com Networks
`The Macintosh network system, Appletalk
`Acorn Computers’ Econet
`
`IBM's NETwork Basic Input/Output system (NETBIOS) aims to provide a fixed
`interface between the operating system and the LAN which is independent of the
`network hardware or software. NETBIOS can be seen as an interface between the
`Presentation and Session Layers of the OSI reference model and it offers both virtual
`circuits and datacgram services. An application communicates with NETBIOS by
`issuing corntnan s in the form of a data structure called a Network Control Block.
`Some PC token ring interface cards use an on-board microprocessor which relieves the
`
`Petitioners‘ Ex. 1017 - Page 10