PC Construction and Maintenance Week 7 Networks. Introduction to networks In computing, the term Network generally means connecting two or more machines.

PC Construction and Maintenance

Week 7

Networks

Introduction to networks

• In computing, the term Network generally means connecting two or more machines together

• Networking covers a wide area of computing.• Networking is becoming increasingly important,

even in the home environment• The cost of networking hardware has dropped

considerably

Why have networks?

• The main reason is to share resources between computers

• Many computers can share a printer

• Internet access can be shared amongst a number of machines

• Access to information and resources shared across a network is often useful in practice

Type of networks

• We are concerned with building a small LAN (Local Area Network) in order to enable file and printer sharing

• Also, a connection to the internet is desirable, for the download of drivers, patches, BIOS upgrades etc

• The word internet literally means between networks

• The internet is really just a large collection of networks, themselves networked together

The theory behind networks

• Network theory could itself encompass a whole course

• We are going to explore the tip of the ice-berg of network theory

• We only need to know the key points in order to have enough knowledge to understand how a small home or business network functions

• There are a just few fundamental facts of computer networking that interest us

Network Theory – Key Points

• All networks are based on a layered model. The ISO 7-layered model is the theoretical ideal model of a computer model

• Real networks are based on a simplified models of the theoretical ideal. But always retain the “layered” quality.

• We are interested primarily in TCP/IP networks. TCP/IP is the protocol that runs the Internet

• We need at the very least to know the essential properties of the layered model in order to configure and troubleshoot basic networks

ISO 7-layer reference model

• 7. Application

• 6. Presentation

• 5. Session

• 4. Transport

• 3. Network

• 2. Data Link (Hardware Interface)

• 1. Physical Hardware Connection

ISO layers explained• Physical Layer:  Transmitting data bits. Voltages, timing

factors, cable standards, etc.Data Link Layer:  Managing basic transmission circuit.  Error detection and correction; message delimitation.Network Layer:  Addressing and routing of packets.Transport Layer:  Establish, maintain, and terminate logical connections. Generating addresses; breaking packets: eliminating duplicate packets; flow control.Session Layer:  Initiating, maintaining, terminate logical session. Presentation Layer:  Display, formatting, encryptionApplication Layer:  Applications.

The TCP/IP model

The Physical Layer

• For ordinary networks, there are two common ways of hooking up several computers

• One way is to use BNC cabling, provided the network cards have a BNC connector on them.

• The most common way at the present time is to use UTP (Unshielded Twisted Pair) cable, in combination with a Hub

BNC based network

• All machines are daisy chained together via BNC cabling and T-pieces

• There is no need for a hub• If a single wire is disconnected, then the whole

network fails• A terminator is plugged in to both ends of the

network• This sort of arrangement is known as “BUS

topology”• A 10-Mbit BNC network is known as 10-BASE2

BUS topology (BNC networks)

UTP based network

• Several machines are connected together via a hub• Several hubs can be chained together to form a larger

networks• For the simple case of a two-computer network, you

don’t need a hub, you can just use a ‘crossover cable’• UTP based networks are fast and reliable• If one wire is disconnected, then only one computer is

affected. The remaining network still functions correctly• UTP networks are known as 10BASE-T, or 100BASE-T,

depending on the speed

Star Topology (UTP based networks)

UTP Cable Wiring

UTP Cable Wiring

UTP Cable Wiring Reference

• 1 White/Orange• 2 Orange• 3 White/Green• 4 Blue• 5 White/Blue• 6 Green• 7 White/Brown• 8 Brown

UTP cables

• A standard UTP network cable is known as a straight cable, or a patch cable

• If the pairs #2 and #3 are swapped with each other, then we have a crossover cable

• Straight cables are the most common, and are used to connect a PC to a hub

• Crossover cables are used to connect a PC to a PC (2 PC network) or , a hub to another hub

Receive/Transmit swapping

• In the 10-base T case. One twisted pair is for receiving a signal (Rx), and the other for transmitting (Tx)

• The hub swaps these over internally so that (Tx) of machine A connects to (Rx) of machine B

• Straight cables do not swap Tx and Rx, crossover cables do.

• If the number of “swaps” between the source and destination machines is odd, then they are wired correctly, else they will not talk to each other

The datalink layer

• The datalink layer deals with transfer of binary data from one network interface card to another

• The dominant datalink protocol used at the current time in a LAN environment is Ethernet

• Some other datalink protocols are still found on LANs (e.g. token ring), but these seem to be dying out

• Because of widespread use of Ethernet equipment, it is now very cheap

Ethernet Networks

• In the Ethernet environment, every network card on the planet has a unique MAC address

• The Media Access Code is used to address different devices on the network

• On a simple network, any device on the network can address any other device within a given subnet using its MAC address

• Devices cannot address other devices on other subnets by using a MAC address

• This is because the datalink protocol is non-routable

The network layer

• IP is the network protocol that drives the Internet (simply IP=InteRnet Protocol)

• It became widely used on UNIX systems to start off with, but has now spread to virtually every other operating system

• IP is needed to ‘surf the web’• Unlike the datalink layer, IP is a routable protocol

and therefore can be used to build large networks (e.g. the Internet)

IP architecture

• In the IP scheme of things, the network as a whole is treated as an interconnected set of subnets

• Each subnet allows communication within itself using the aforementioned datalink protocol

• Subnet A is connected to Subnet B via a device known as a router

• Each machine on the network (even on the public internet) must have a unique IP address

IP addresses

• IP addresses have a four byte form and are normally written using a dotted notation

• For example 150.204.51.254• One some networks, static I.P addresses are used,

which means that each machine handles it’s own IP configuration locally

• Some networks employ protocols to allow an IP address to be assigned dynamically at boot time

Configuring IP under windows 95

• As a test case, we will go through the steps of configuring IP networking under windows 95

• Firstly, we need to have a network adapter available

• We need to check using device manager that a network card is installed

Installing the network card driver• If windows has automatically recognised the card,

then we will see it in device manager• If not, we may need to use the “add hardware wizard”• Sometimes, windows can see the card, but doesn’t

know what type it is. In this case, it will show up as “Unknown device”, next to a question mark symbol

• If this happens, remove the unknown device from device manager and reboot the PC

• Windows should then properly detect the card on the next boot

Installing the network components

• Once we have the card install correctly, we need to add the TCP/IP component

• Right-click on Network-Neighbourhood (Or select network) in Control Panel, and then click on “Add” to add a protocol (You will need the windows CD in the drive)

• The protocol is Microsoft -> TCP/IP

TCP/IP Parameters

• After adding the TCP/IP protocol, we need to configure the relevant settings to run the machines on our network

• The main setting is IP address (I have reserved a separate address for each group, written on a list)

• The other settings are Subnet Mask,Gateway, DNS servers and WINS servers

Subnet mask

• For our network, the subnet mask is always 255.255.252.0

• The subnet mask defines how we choose to split our physical subnet into logical subnets

• The non-zero (left hand portion) defines the network part of the address

• The zero portion (right hand portion) defines the host part of the address

Gateway

• For our network, the address of the gateway (a.k.a the router) is 150.204.51.254

• The router is just a box that bridges the gap between our network and the outside world

• The network could work as a self contained entity without a router (e.g. as in a home network)

• The router is needed to access the internet

DNS servers

• DNS servers are just other computers on our network that translate things like web-site names into IP addresses

• Without them, we could theoretically still browse the internet, but we would have to refer to websites by number, rather than by name

• The address of the three DNS servers on our network are 150.204.51.1 150.204.51.5 150.204.51.17

WINS Servers

• WINS servers are used to translate between Microsoft networking names and IP addresses

• We need to configure WINS servers to allow are machines to use Microsoft network resources such as shared printers and files

• Our WINS servers are 150.204.13.108 and 150.204.40.1

Completing the configuration

• Any other protocols on the system, such as NetBUI and IPX should be removed on PCs on our network

• After rebooting, the machine should be enabled for TCP/IP

• This can be tested by launching Internet explorer• If that fails, then try “ping 150.204.51.254” to

bounce a packet of data off the router from the DOS prompt.

Notes on home networking• In the home environment, there are typically no WINS

or DNS servers available• In that case, it is still possible to run TCP/IP over the

network• Its just that local file configurations are used instead of

network servers• Local file configuration becomes impractical when

working with a large number of machines though• Microsoft’s custom protocols may sometimes be more

suited to a small home network (but not for internet sharing)