Benefits of a LANThere are numerous benefits for having a LAN. Each network benefit is important and even critical to an organization's success. These benefitsinclude:

Resource sharing

Workgroup synergy

Management

Centralized

Decentralized Data access and integration

Economic benefits 1

LAN Architecture

Software An end-user appplication may use a software protocol suite such as the TCP/IP or ISO/OSI

Hardware The physical network medium designed to carry informational signals, such as coaxial, twisted-pair cable, of fibre-optical material carrying multiband moduated laser light.

2

LAN Architecture Topology

A network constructed of the afore mention materials on one or more interconnecting plans.

Backbone (bus topology)This has been the typical LAN topology for Ethernet since its inception.

This configuration has one large co-axial cable running thoughout the area.

3

IBM CompatibleIBM Compatible IBM Compatible

IBM Compatible IBM Compatible

LAN Architecture Star configuration

This method uses a central location or hub from which a number of signal carrying cables goes out to each individual device on the branch of the LAN. A single hub can serve from 2 to 128 Lan devices (maybe more) on a single network domain. One otheradvantage to the star configuration is that the maximum distance between any two nodes is always only two segments long.

4

IBM Compatible

IBM Compatible

IBM Compatible

IBM Compatible IBM Compatible

IBM Compatible

Workstation Workstation

Workstation

Workstation

LAN Architecture Ring configuration

A true configuration has two ports on every node: one for input and one for output.

5

Token-ring

Workstation

Workstation

Workstation

Workstation

Workstation

Workstation

Workstation

Workstation

LAN Methodologies

Ethernet - IEEE 802.3 Ethernet is assumed to be the LAN method unless other-wise stated. More than 85% of all installed networksconnections were Ethernet by the end of 1996. This means there are over 150 million interconnected workstations, PC and servers using Ethernet today.

ATM (Asynchronous Transfer Mode)ATM eliminates inefficiencies by dynamically sharing network bandwidth among multiple logical connections.ATM has been defined at speeds of 45 Mbps, 100 Mbps,155 and up to 622 Mbps. 6

LAN Methodologies ATM (Continued)

The ATM LAN equipment includes ATM switches,routers, hubs, bridges, and workstations.

Token Ring - IEEE 802.5The Token Ring network was originally developed by IBM. It is still IBM's primary local-area network (LAN)technology. Token-passing networks move a small command frame, called a token around the (circular ring)network. Possession of the token grants the possessor the right to transmit data. To transmit data, the tokenthe token is changed to a data frame and the informationis attached. 7

LAN Methodologies

FDDI - (Fiber Distrubuted Data Interface) FDDI is frequently used as a backbone technology as well as a means to connect high-speed computers in alocal area.

ISO (International Organization for Standardization) hascreated an international standand for FDDI. FDDI specifies a 100-Mbps, token-passing, dual-ring LAN using a fiber-optic transmission medium.

8

LAN Methodologies FDDI - (Continued)

The dual-ring fiber-optic medium allows for true bi-directional simultaneous full duplex operation at 100Mbps on each fiber channel.

FDDI uses optical fiber as a transmission medium. Optical fiber offers several advantages over traditional copper wiring:

Security (fiber does not emit electrical signals that can be trapped) Reliablility (fiber is immune to electrical interface) Speed (optical fiber has much higher throughput potential than

copper cable) No interference from the outside EMI (electromagnetic interference) source. Does not produce any EME interfere with and other communication medium (also a security advantage) 9

LAN Components

Topology - A description of the physical construct or layout of a network.

Backbone - The primary connectivity mechanism of an Ethernet network. All systems that have connectivity on the backbone can

have connectivity to each other.

Segment - A continuous length of cable commonly joined with other segments.

Repeater - A device that amplifies and regenerates the data signal bit by bit in order to extend the distance of the transmission. A repeater does not read or interpet the data.

10

LAN Components Bridge - A device that connects two or more networks segments of the same physical media type. A bridge examines the hardware address fields of a network packet and the filters

based on addresses from one network segment to another and vice versa.

Router - A device that has two or more network interfaces. It examines the software proto-col (IP) address, selects an appropriate travel path and forwards the packet accordingly between seperate networks. Routers usually forward packets belonging to a single protocolfamily.

Gateway - A device that interconnects two or more communications networks based on different protocol suites. The gateway performs any necessary protocol conversions.

Switch - A multiport device which minimally functions as a bridge, and provides for the logical dynamic connection and disconnection between any two cable segments without

operator intervention. The switch is a high-speed device because multiple data paths canbe established and used simultaneously.

Concentrator - The central device through which all hosts in a twisted pair Ethernet installation are connected.

Hub - A central device through which various types of network packets can flow. 11

Networking Models

The three networking models that provide a frameworkfor network communication are:

ISO/OSI reference model

TCP/IP suite (TCP/IP model or TCP/IP)

SPX/IPX (Sequenced Packet Exchange/ InternetworkPacket Exchange).

12

Networking Models

SPX/IPX

This is a propriety protocol operating on Nevell Netware local area networks. SPX/IPX uses most Ethernet network interfaces such as IEEE 802.2,802.3 and Ethernet-II. SPX/IPX is not supported on the internet. If informationdestination are systems connected top the internet, the model must be convertedto TCP/IP prior to making the connection to the internet.

For a through treatment of integration SPX/IPX into TCP/IP environment, referto:

Network Personal Computers With Tcp/IP, Craig Hunt, O'Rielly, 1995

Novell's Guide to Integrating UNIX and Netware Networks, Novell Press,1994

13

Networking ModelsThe ISO/OSI reference model uses seven layers to describe its networkcommunications framework. Brief description of each layer are as follows:

Application layer Consists of user-accessed application programsand network sevices.

Presentation layer Defines the way in which cooperating networksrepresent data.

Session layer Manages the connections between cooperatingapplications.

Transport layer Responsible for end-to-end messaging from oneapplication program to another, also knows as end-to-end communication.

Network layer Manages data addressing and delivery betweennetworks. This layer fragments data into smallerpieces that the data link layer can handle.

Data link layer Manages the delivery of data across the physicalnetwork. This layer provides error detection andpacket framing.

Physical layer Describes network hardware, including electricsignal characteristics such as voltage and current.

14

Networking Models

15

The TCP/IP ModelThe TCP/IP model describes its network communications frame-work usign five layers. Brief descriptions of each layer are as follows:

Application Layer Consists of user-accessed application programs and network services. This layer is also responsible for defining the way in which cooperating networks represent data. A gateway functions at this layer.

Transport Layer Manages the transfer of data using unacknowledged transport protocols. This layer also manages the connections between cooperating applications.

Internet Layer Manages data addressing and deliverybetween networks, as well as fragmentingdata for the network interface layer. A router functions at this layer.

Network Layer Manages the delivery of data across the physical network. A bridge works at this level.

Physical Layer Describes the network hardware, including electical signal characteristics such as voltage and current. A repeater is at this layer

Ethernet Definition

16

Ethernet is a packet-switching network that is based on a 10 Mbytes/sec broadcast technology with distri-buted access control.

Ethernet is composed of three major elements:

Hardware I/O that transfers data to and from apacket switching network of computers

A packet switching network allows concurrentmultiple communications among nodes on a network.

The Ethernet packet which is a unit of data sent across a network

This unit of data includes the destination address, the source address, and a form of parity checking all wrapped around a block of information.

The Ethernet access method protocol (CSMA/CD),which is used to control packet transmission and flow over the Ethernet hardware.

17

CSMA/CDHosts send messages on an Ethernet LANusing a Network Interface Layer protocoland CSMA/CD.

Host hasmessage

Traffic onnetwork?

sendmessage

anycollision

wait, backoff exp( )

No

Yes success

No

Yes

Multiple access

Carrier sense

Collision detect

18

Switched Ethenet

Switched Ethernet reduces the number ofcollisions on a network by utilizing an intelligent buffering system and control ofthe backbone. By removing the physical backbone network wire and replacing it witha central hub device that can receive, store, and transmit packets.

19

Fast Ethenet

Fast Ethernet is capable of burst rates upto 100 Mbytes/sec over a single duplexconnection and over 200 Mbytes/sec on afull-duplex dual-ported configuration. In a dual-ported full-duplex configuration, data is always traveling one direction on a singlecable from point A to point B and coming back on a different cable from point B to point A.

20

Multimode EthenetBecause of packet-switching and the Ethernet being so versatile and adaptable, an Ethernet network can incude a variety of network speeds and interfaces. Mixing products from various vendors is also possible because of the standards laid out in the IEEE 802.3 specifications. For example:

HUB

WorkstationWorkstation

Workstation

Workstation

WorkstationWorkstation

Workstation

WorkstationWorkstation

Workstation

Token-ring

WorkstationWorkstation

Workstation

Workstation

Workstation10 Base TFiber

100 Base TFiber

Fiber

1-Gbytes/sec Ethernet

Server

21

Ethenet Address

An Ethernet address is a host's unique hardware address. It is 48 bits long and is displayed as 12 hexadecimal digits (6 groups of 2 digits),separated by colons (:). An example of a complete Ethernet address is8:0:20:1e:56:7D.

Unique Ethernet addresses are administered by Xerox. The first threeoctets are vendor-specific and are designated by Xerox. Sun systemsalways begin with the sequence 8:0:20. Sun assigns the last three octetsto Sun products it manufactures. This method ensures that each node onan Ethernet has a unique Ethernet address.

Sending Messages

There are two types of ethernet addresses which can be used to communicate across the network:

Unicast address

A host sends a message to another host on the Ethernet using a unicast address. Individual host Ethernet addresses are used for one-to-one, unicast transmissions.

Broadcast address

A hsot sends a message to all hosts on the Ethernet using a broadcastaddress. The Ethernet broadcast address is all one (ff:ff:ff:ff:ff:ff inhex). When an Ethernet frame is received with a destination addressof all ones, the Network Interface layer passes it to the next layer.

22

Ethenet Frame Analysis

An Ethernet frame is a single unit of data transported through the LAN. A frame is a series of bits with a definitebeginning and end. The Ethernet specification describeshow bits are encoded on the cable and how hosts on the network detect the beginning and end of a transmission. Thefollowing diagram illustrates the relationship in this analogy:

Preamble 64 bits

D addr48 bits

S addr48 bits

Type16 bits

Data(maximum 1500 bytes)

CRC32 bits

Octet location: 1-6 7-12 13-14 15-1514(the max) last 4 octet

Host in an Ethernet LAN use this information to receive andtransmit data.

PrembleThe 64-bit Ethernet preamble field, composed of ones and zeros, is used for synchronization. Synchronization helps thenetwork interface determine where an Ethernet frame begins.

Destination AddressThe destination address field is the Ethernet address of the destination host.

23

Ethenet Frame (Continued)

Source Address

The source address field is the Ethernet address of the sending host.

Type

The fourth field of the Ethernet frame describes the type of dataencapsulated in the Ethernet frame (such as IP, ICMP, AddressResolution Protocol (ARP) or Reverse ARP (RARP)).

Data

The data field holds a minimum of 46 bytes and a maximum of 1500 bytes of information. The network hardware determinesthe maximum number of octets of data in a frame. This is calledthe MTU (maximum transfer unit). If the data to be transmittedis less than 46 bytes, the data is padded with zeros to reach to reach the 46-byte minimum.

Cyclical Redundancy Check (CRC)

The CRC field is used for error detection. The value is calculatedbased on frame contents, by the sending host. The receiving hostuses the same algorithm to recalculate the CRC upon arrival, andthen compares it with the frame CRC value. If the two values arenot the same, the frame is ignored.

24

Ethenet Frame Encapsulation

Application Data

Application Data

Application Data

Application Data

Transport header

Internet header

Ethernet Header

25

Useful Troubleshooting Commands

Snoop

Location - /usr/sbin/snoop

You can use snoop to capture network packets and displaytheir contents. Packets can be displayed as soon as they arerecieved, or saved to a file.

26

Useful Troubleshooting Commands

Netstat

Location - /usr/bin/netstat

Use the netstat - i command to show the state of the Ethernetinterfaces.

# netstat - i

Name - The name of the device (interface). Mtu - The maximum transfer rate in bytes. Net/Dest - The network number. This field references the file /etc/inet/networks. This file is discussed later. Address - The IP address for that interface. Ipkts/Ierrs - Shows the input packets and errors. Opkts/Oerrs - Shows the output packets and erros. Coll - The number of collisions on this interface. Queue - The number of packets awaiting transmission.

To display the contents of the routing table for the local system, use the netstat - r. command

27

Useful Troubleshooting Commands

Ifconfig

Location - /usr/sbin/ifconfig

The ifconfig command is used to display information aboutthe configuration of the network interface specified. The following example shows the configuration of a 100BaseT interface, including its IP and Ethernet addresses.

# ifconfig hme0hme0: flags=863<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICAST> mtu 1500 inet 128.235.35.124netmask ffffff00 broadcast 128.235.35.255 ether 8:0:20:80:d0:a7