Chapter 14

Making Network Connections

Home | Chapters | Homework | Grades | Final Project | Calendar | Forums | Syllabus

Forum Policy | Assignment Policy | Games | Tests | E-Mail Mr Hull | FTP

Main | Chapter 1 | Chapter 2 | Chapter 3 | Chapter 4 | Chapter 5 | Chapter 6 | Chapter 7

Chapter 8 | Chapter 9 | Chapter 10 | Chapter 11 | Chapter 12 | Chapter 13 | Chapter 14 | Chapter 15

LAN-to-LAN Connections

Once we connect two LAN's together we are using a WAN.  But why would we need to connect multiple LAN's together to create a WAN?  We learned in Chapter's 1 and 13 that we might need to do this because the distance between two LAN's is farther then the maximum length of LAN media.  But this is not always the case.  One reason to have multiple LAN's connected might be to separate users.  For example in a school you could have one LAN for students and one for administrators and staff.  This can prevent the students from gaining access to the guidance database.  You could allow teachers access to the admin LAN so they can post there grades, and attendance etc. 

Another reason to connect to LAN's might be the way the LAN's were created.  As computers were introduced to the workplace they were standalone stations.  As more were added they might have hooked there computers together to share information with each other in there department.  As time progressed you might find yourself with multiple small departmentalized LAN's that you need to hook together.  In this case you might want to rely on WAN technologies to connect the LAN's.

You also might want to split a large LAN into smaller LAN's to improve performance.

Repeaters, Bridges, Routers, and Gateways

  • A repeater is a hardware device that repeats the signal to allow for longer cable runs.  They operate at the physical layer, and do not segment the network.

  • A bridge is a hardware device used to connect two LANs together.  It operates at the data link layer and can connect LAN's with the same media type, or different media types.  One thing about earlier bridges is they connected networks using the same MAC protocol.  Newer bridges called brouters can connect networks with different MAC protocols.  So you can use a brouter to connect Token Passing and CSMA/CD networks.  Also older bridges used to forward all packets across, newer ones called learning bridges builds a routing table to determine if data needs to pass through it or not.  To build the routing table it can use either Spanning Tree Algorithm or Source Routing.   With Spanning Tree Algorithm the bridge will build its table by looking at the source address on messages as they come in and adding them to the table.  If it doesn't have an address that a packet is destine for it will flood the network requesting it's location.  Once found it adds it to the table.  In Source Routing the sending host sends out a discovery packet to locate the destination host and lets the bridge know the route.

OSI Model - This shows were each of the hardware devices reside in the OSI model.

  • A router is used to connect LAN's, it operates at the network layer.  Because of this a TCP/IP router cannot route a SPX/IPX packet.  The network layer is responsible for end to end routing, the router will look at the destination address and determine if itself is the destination.  If it isn't it will look at it's routing tables and determine where it needs to send the data to get it to it's destination.  It will then pass it along to the data link layer where physical addressing information will be added then to the physical layer where it will be sent out onto the medium. 

  • A gateway is a protocol converter.

Switches and Virtual LANs

There is more of a drive today to have faster and faster LAN's.  This is because the applications that we use are starting to require more bandwidth and more people are starting to use the network.  We want to increase the speed of our network with the lowest cost.  To do this we can try and reuse the existing lines and upgrade certain components on the network.  For example we might replace an older 10 Mbps hubs with faster 100 Mbps switches.  This upgrade helps us in more ways then one, when we move to the switch you now have a full duplex network, you have the higher speed, and you have a single collision domain.  A switch or a smart hub only sends the data out the port that contains the destination node.  Because of this it is possible to have two separate communications.  This will reduce collisions.  A switch will move data in one of two ways, one way is cut-through  (aka cross-point switches) and the other is store and forward.  A Cut-through will read the MAC header and just pull out the destination information and forward it to the appropriate port.  There is no error detection in this method.  A Store and Forward switch is essentially a high speed router.  It will store the data and delete it once it has received an acknowledgment from the destination next node.
Virtual LAN's allow you to segregate your LAN at the switch level.  There are three types of separation you can do.  At the Physical Layer of the OSI model the port you plug the computer in the switch will determine  which network the node is on.  On a 24 port switch you can have port 1-12 be on VLAN0 and ports 13-14 be on VLAN1.  At the Data Link Layer the MAC address will determine the VLAN membership.  At the Network Layer level the IP address will determine the VLAN membership

Asynchronous Transfer Mode (ATM) is a high speed switching technology used to provide services for LAN's and WAN's.  Because of it's high speed it is usually used for backbone, and high-bandwidth networks.  The typical speed of an ATM is 155 Mbps but is scalable and capability of faster speeds.  The data in ATM is broken down into small fixed sized frames called cells.  A cell is 53 bytes long and carries 47 or 48 bytes of data.  The remaining 5 or 6 bytes are used for the head which contains addressing, error-checking, cell type identification, and flow control.  

LAN-to-Host Connections

Suppose you have a network that was using a mainframe computer and all the users used dumb terminals to connect to the mainframe and run there application.  As personal computers start to populate the desktops and the dumb terminals are being replaced you still need a way to access the old mainframe.  In this section we will take a look at some of these methods, we will be referring to the mainframe as a host

  • The Host as a LAN node - Some host's have the ability to hook up to a LAN and become a node on the network.  In this case the host acts as a server.  This is the easiest way to do this.  But this is not always the case.

  • Asynchronous Connections - Every computer has the ability to communicate using an asynchronous data link protocol.  The communication ports on a computer communicate asynchronously.  This communication can be as a file transfer or a terminal emulation.

  • Dedicated Connection per Microcomputer - Some host's can support multiple asynchronous connections.  In this case each node that needs access can have a direct connection to the host.

  • Multiplexer - This hardware device allows multiple users to connect to one communication channel. 

  • Shared Asynchronous Connections - You would have a communications server connect to the host and have the other nodes connect to the server.  This was the only dedicated connections are between the server and the host.  All the other nodes on the network share those connections.

  • Other Types of Host Connections - Other types of connections that might exist are through terminal emulations.  IBM-3270 or LU 6.2 Connections for example. 

Interconnection Utilities

Some of the utilities that you might use to establish connections or utilities that would assist in networking tasks are described here.

  • File Transfer Utilities - FTP, Kermit, XModem, YModem, and ZModem are all utilities that allow you to transfer files over a network.  Most of these would be found in a terminal emulation program.

  • Remote Logon - Remote logon software allows a remote node to act as a local node.

  • Remote Access - Remote access allows a user to work in a remote location and have access to network resources.  This can occur in one of two ways, remote control and remote node.  Remote Control allows the remote users to see the environment of the local node.  Keyboard commands and mouse movements are sent to the local node and processed, the local node then returns updates information which is displayed on the remote nodes screen.  With remote node the remote computer acts as a node on the local network.  Remote node is usually slower because more information must be passed back and forth. 

More Information


Click Here to download the slides for this chapter

(NOTE: You must have PowerPoint or PowerPoint Viewer if you don't have  either Click Here to download PowerPoint Viewer.)

Home | Chapters | Homework | Grades | Final Project | Calendar | Forums | Syllabus

Forum Policy | Assignment Policy | Games | Tests | E-Mail Mr Hull | FTP