How To Design A Wide Area Network

WAN Overview:

WANs are usually required for high volume, long-distance data traffic.

To implement WANs, you can use the following transmission media:

• PSTN (Public switched telephone network)
• High-speed, high-bandwidth dedicated leased circuits
• High speed fiber-optic cable
• Microwave transmission links
• Satellite links
• Wireless radiated media (radio frequency)
• The internet

There are two types of WANs:

• Enterprise networks: When a network connects a company’s branch offices and divisions, it becomes an enterprise- wide network. For example, a corporation may have sites on every continent, all of which are interconnected to form one wide area network.
• Global networks: When a network spans several countries and continents and includes many types of organizations and individuals, it can be labeled  global. These networks serve multinational corporations and scientific, academic and military establishments. The internet, often called the “network of networks” fits that definition.

Public Network Services:

Two of the most popular methods are the PSTN and the internet.

PSTN:

•  PSTNs were originally designed exclusively for telephones but have become highly sophisticated, able to handle different kinds of data transmission, including digital data transmission.
• The PSTN provides a number of options for data transmissions, including services that route packets between different sites.

Available services and some possible transmission rates include the following:

             Service                                            Transmission Rate

              Switched 56                                    56Kbps

              X.25                                               56Kbps

              T1 circuits                                       1.544Mbps

              T3 circuits                                       44.736Mbps

              Frame relay                                     1.544Mbps

              SMDS                                            1.544Mbps

              ISDN                                              1.544Mbps

              ATM                                              44.736Mbps

The Internet:

It is a shared network of government agencies, educational institutions, private organizations and individuals from over a hundred nations.

No one owns the Internet; anyone can have access to the transmission media.

SLIP and PPP:

• Serial Line Internet Protocol (SLIP) and Point-To-Point (PPP) are two very common protocols used to transmit IP packets over serial line and telephone connections, most often as part of a dial-up internet connection.
• SLIP an PPP are similar protocols. But some differences are there:
• Basically PPP is a multistage protocol signal transport mechanism. SLIP is designed for maintaining one type of traffic protocol  (TCP/IP traffic) at a moment.
• With SLIP, you must know both the IP address assigned to you by an ISP and the IP address of the remote system your computer will be dialing into. PPP deals with these problems by negotiating configuration parameters at the beginning of your connection.
• It also said that PPP can be  negotiate header compression. SLIP does not offer this type of compression.
• Another things about PPP that it offers IP enhanced security  protection.

Switching:

• Switching is an important technique that can determine how connections are made and how data movement is handled on a WAN.
• Data sent across the PSTN or other internetworks can travel along different paths from sender to receiver.

Three major switching techniques are:

• Circuit switching
• Message switching
• Packet switching

Circuit switching:

• In circuit switching, a dedicated physical connection is established between the sender and the receiver and maintained for the entire conversation.
• PSTN uses a circuit switching system.
• Before any two computers can transfer data, a dedicated circuit must be established between the two.
• The sending machine requests a connection to the destination, after which the destination machine signals that it is ready to accept data.
• The data is then sent from the source to the destination and the destination sends acknowledgements back to the source.
• When the conversation is finished, the source sends a signal to the destination, indicating that the connection is no longer needed and disconnect itself.
• A advantage of circuit switching is that the dedicated transmission channel the machine establish provides a guaranteed data rate.
• Since the channel is always available, it does not need to be requested again.
• Disadvantages: It is often an inefficient use of the transmission media.
• Dedicated channels require more bandwidth than non-dedicated channels, so transmission media can be expensive.
• Also, this method can be subject to long connection delays; it may take several seconds to establish the connection.

Message switching:

• Message switching is unlike circuit switching in that it does not establish a dedicated path between two communicating devices.
• Instead, each message is treated as an independent unit and includes its own destination and source addresses.
• Each complete message is then transmitted from device to device through the internetwork.
• Each intermediate device receives the message, stores it until the next device is ready to receive it and then forwards it to the next device.
• For this reason, a message switching network is sometimes referred to as a store-and-forward network.

Advantages:

• It provides efficient traffic management.
• It reduces network traffic congestion. The intermediate devices are able to store messages until a communications channel becomes available.
• Its use of data channels is more efficient than circuit switching.
• It provides asynchronous communication across time zones.

Disadvantages:

• The delay introduced by storing and forwarding complete messages make message switching unsuitable for real-time applications.
• It can be costly to equip intermediate devices with enough storage capacity to store potentially long message.

Packet switching:

• In packet switching, message are broken up into packets, each of which includes a header with source, destination and intermediate node address information.
• Individual packets don’t always follow the same route; this is called independent routing.

Independent routing offers two advantages:

Bandwidth can be managed by splitting data onto different routes in a busy circuit

If a certain link in the network goes down during the transmission, the remaining packets can be sent through another route.

Methods of packet switching:

1. Datagram packet switching
2. Virtual-circuit packet switching

Datagram packet switching: In a datagram packet-switched network, a message is divided into a stream of packets.

Each packet is separately addressed and treated as an independent unit with its own control instructions

The switching devices route each packet independently through the network, with each intermediate node determining the packet’s next route segment.

Internetwork Connectivity Devices for Networking

Internetwork Connectivity Devices

An internetwork connectivity consists of two or more independent networks that are connected and yet maintain independent identities. An internetwork may include different types of networks. To connect independent networks, we use internetwork connectivity devices. The devices are: routers, brouters, gateways, CSUs/DSUs etc. 

Some of the benefits of internetworking are:

• Reduces network traffic:With internetwork connectivity devices, most traffic stays on the local network and only packets destined for other networks cross internetwork connectivity devices.
• Optimizes performance:The benefit of reduced traffic is optimized performance.
• Simplifies management:Network problems can be more easily identified and isolated in smaller networks, as opposed to one large network.
• Efficiently spans long geographical distances: Because WAN links are many times slower and more expensive than LAN links, having a single large network spanning long distances can complicate network management and slow network performance. We can more efficiently span long distance by connecting multiple smaller networks.
• Interconnection: To connect our network with those of other organizations is also a good reason.

Routers:

• Practically Routers are devices that connect two or more networks. 
• Router consist of a combination of hardware and software. 
• In Router the hardware can be a network server, a separate computer or a special black box device.
• For Router the two main pieces of software is the OS and the routing protocol. 
• Management software is another component of router. 
•  For routing the Routers use logical and physical addressing to connect two or more logically separate networks. 
• Different Router accomplish this connection by organizing the large network into logical network segments (subnet).
• In each of these sub network it has given a logical address. 
• For packet transferring each packet in addition to having a physical device address, has logical network address. 
• The network address allows routers to more accurately and efficiently calculate the optimal path to a workstation or computer.
•  Routers keep the networks separate and router processing is generally slower than bridge processing.
• Routers are more intelligent than bridges because they use algorithms to determine the best path to send a packet to a network. 
• By passing packets only according to network addresses, routers can help prevent a broadcast storm.
• Routers list network addresses in routing tables.
• These tables contain all known network addresses and possible paths. 

Terms to quantify the routing cost:

• Hop count describes the number of routers a message must pass through to reach its destination.
• Tick count describes the amount of time required for a message to reach its destination. A tick is 1/18 second.
• Relative expense is a number you can assign based on the actual monetary cost or some other relevant criteria required to use a given link. 
• Route discovery is the process of finding the possible routes through the internetwork and then building routing tables to store that information.

Two methods of route discovery are:

• Distance-vector  
• Link-state

Distance-Vector Routing:

• In distance-vector routing, each router advertises its presence to other routers on the network.
• Periodically each router on the network broadcasts the information contained in its routing table.
• The other routers then update their routing tables with the broadcast information they receive.
• These periodic broadcasts of routing table information by the routers performing distance-vector route discovery add up to a noticeable amount of traffic.
• This traffic is not a problem in LANs, because plenty of bandwidth is available and the number of routers is usually low.
• However, it seriously affect performance in a WAN.
• In a large internetwork, distance-vector routing tends to be quite inefficient as route changes must be broadcast through the network from router to router and because changes are contained within complete routing tables, it can take a long time before all the routers on the network know of a change.

Link-State Routing:

• Link-state routers broadcast their complete routing tables only at startup and at certain intervals-much less frequently than distance-vector broadcasts.
• This type of routing generates less network traffic than the distance-vector method.
• The major difference between the link-state and distance-vector methods is that once the initial routing-table exchange has occurred, a link-state router will generally broadcast routing updates only when it detects a change in its routing table. When it does broadcast, it sends only information about the change, it doesn't send its complete routing table.
• Selection of the optimum route can be dynamic or static.
• Dynamic route selection permits routers to constantly adjust to changing network conditions.
• With static route selection, packets must always follow a predetermined path.
• Dynamic route selection uses the cost information that is continually being generated by routing algorithms and placed in routing tables to select the best for each packet.
• Other routers that receive broadcast messages regarding changes in the state of network links use this information to update their own routing tables.
• As only the changes are sent, these updates can be done in less time.
• Once a router has created its routing table, it can use the cost information contained within that table to calculate the best path through the internetwork.
• Routing protocol can select the best path based on the minimum number of hops, number of ticks or relative expense.
• As network conditions change, the router can select different paths to maintain the lowest possible costs.
• The router can even select new paths “on the fly” as it is transmitting packets.
• If changes occur during a transmission that make one route suddenly less attractive than another, the router can send the remaining packets of the transmission along a different path from the packets in the first part of the transmission.
• With static route selection, the data path is not selected on the fly by the routers involved.
• Instead, the data path is designed in advance.
• Either the network administrator or a computer on the network selects a route for the data from a predefined table.
• All packets are then forced along that route and intermediate routers are not allowed to make route selection decisions.
• Static route selection tends to be less efficient than dynamic route selection because it cannot adapt to changing network conditions.

Brouter:

• A brouter is a router that can also bride.
• A brouter first tries to deliver the packet based on network protocol information.
• If the brouter does not support the protocol the packet is using or cannot deliver the packet based on protocol information, it bridges the packet using the physical address.
• True routers simply discard a packet if it doesn’t have a correct logical address.
• A brouter can be a more affordable option to having both a router and a bridge.
• Keep the following in mind when working with routers:
• Some routers may not follow standards. This can cause problems when we use different vendors’ routers on the same network.
• Be sure the router is rated to handle the speed of the network connection.
• Routers slow down network communications to a small extent, so don’t use them unnecessarily.
•  Routable protocol: DECnet, DDP, TCP/IP, NWLink IPX, OSI, XNS
• Non-routable protocol: LAT, NetBEUI
• A gateway is a device that can interpret and translate the different protocols that are used on two distinct networks.
• Gateway can be comprised of software, dedicated hardware or a combination of both.
• Gateways can function at network layer.
• When you need to have different environments communicating, you may wish to consider a gateway.
• A gateway can actually convert data so that it works with an application on a computer on the other side of the gateway.
• You can connect systems with different communication protocols, languages and architecture using a gateway.
• Gateway can be slow because they need to perform intensive conversion and they can be expensive.

CSUs/DSUs:

• Sometimes, when expanding your network, it is less costly and easier to use existing public networks, such as the public telephone network in your area.
• Connecting to some of these networks requires the use of CSUs/DSUs (channel service units/digital service units).
• Network service providers may require you to use a CSU/DSU to translate in signals of your LAN into a different signal format and strength for use on their transmission media.
• CSUs/DSUs are also useful for shielding your network from both noise and voltage currents that can come through the public network.

How To Expanding A Networks

Network Connectivity Devices:  

To expand a single network without breaking it into new parts or connecting it to other networks, we can use the following devices:

•  Hubs
•  Repeaters
•  Bridges
•  Multiplexers

Hub:

• All networks require a central location to bring media segments together
• The central locations are called hubs
•  A hub organizes the cables and relays signals to the other media segments.
• Important things about hub:
• There is a limit to the number of hubs that can be connected to each other to extend a network.
• When possible, connect each hub directly to a server network card rather than to another hub.
• Label the connection on the hub.
• The more hubs data passes through, the slower the connection.
• Passive hubs
• A passive hub simply combines the signals of network segments.
• There is no signal processing or regeneration.
• As it does not boost the signal and absorb some of the signal, it reduces by half of maximum cabling distance permitted.

Here, each computer receives the signals sent from all the other computers connected to the hub.

Active hub:

• It regenerates or amplify signals.
• The distance between devices can be increased.
• They also amplify the noise as well.
• They are more expensive than passive hub.
• Bcz some active hubs function as repeaters, they are sometimes called multi-port repeaters.

Intelligent hub:

Intelligent hubs can regenerate signals but it can perform some network management and intelligent path selection.

A switching hub chooses only the port of the device where the signal needs to go, rather than sending the signal along all paths.

Many switching hubs can choose which alternative path will be the quickest and send the signal that way.

Repeaters:

• All transmission media attenuate the electromagnetic waves that travel through them.
• Adding a device that amplifies the signal can allow it to travel farther, increasing the size of the network.
• Devices that amplify the signals in this way are called repeaters.
• Repeaters fall into two categories: amplifier and signal-regenerating repeaters.
• Amplifiers simply amplify the entire incoming signal and amplify both signal and noise.
• Signal regenerating repeaters create an exact duplicate of incoming data by identifying it amidst the noise, reconstructing it and retransmitting only the desired information.
• The original signal is duplicated, boosted to its original strength and sent.

Bridges:

Bridges connect network segments.

The use of bridge increases the maximum possible size of your network.

A bridge selectively determines the appropriate segment to which it should pass a signal.

It does this by reading the address of all the signals it receives.

The bridge reads the physical location of the source and destination computers from this address.

The process works like:

A bridge receives all the signals from both segment A and segment B.

The bridge reads the addresses and discards all signals from segment A that are addressed to segment A, bcz they do not need to cross the bridge.

Signals from segment A addressed to a computer on segment B are retransmitted to segment B.

The signals from segment B are treated in the same way

Through address filtering, bridges can divide busy networks into segments and reduce network traffic.

Network traffic will be reduced if most signals are addressed to the same segment and do not cross the bridge.

Two types of bridges:

• Transparent bridges keep a table of addresses in memory to determine where to send data
• Source-routing bridges require the entire route to be included in the transmission and do not route packets intelligently.
• Multiplexing allows you to use more bandwidth of the medium by combining two or more separate signals and transmitting them together.
• The original signals can then be extracted at the other end of the medium. This is demultiplexing.
• Multiplexing provides a way of sharing a single medium segment by combining several channels for transmission over that segment.

Three major methods of multiplexing are:

1. Frequency division 

2. Time division 

3. Statistical time division

Frequency-Division multiplexing:

FDM uses separate frequencies to combine multiple data channels onto a broadband medium.

You can use FDM to separate traffic traveling in different directions in a broadband LAN.

Time-division multiplexing:

• TDM divides channel into time slots.
• Each of the devices communicating over this multiplexed line is allocated a time slot in a round-robin fashion.
• If a device does not use its time slot, that slot is wasted.

Statistical Time-division multiplexing:

• TDM systems can be inefficient if many slot times are wasted.
• StatTDM provides an intelligent solution to this problem by dynamically allocating time slots to devices on a first-come, first-serve basis.
• The number of time slots allocated to a particular device depends on how busy it is.
• You can use priorities to allow on device greater access to time slots than another.
• For the multiplexer on the receiving end to determine which signal a particular time slot is carrying, there must be a control field tat identifies the owner attached to the data.

Designing the Local Area Network

Network Scale:

• How many clients do you have?
• How far apart are the computers?
• What software are you using?
• What software will you use?
• What special requirements do you have?
• How much can you spend?

How many computers do you have?

• The number of client computers you have is the most important factor in network design
• You all other design factors are affected by the size of your network

Peer Network (2-10 Users):

• A peer network provides basic connectivity between computers but does not set apart any central computer as a server or provide many of the security features of a centralized client server network.
• If you have only a few users and security is not a major concern, consider suing a peer network.

Peer networks are good for:

-File sharing

-Printer sharing

-E-mail

-Tight budget

-Easy installation

 They are not good for:

               -Security

               -Backup

               -Organization of data

               -Database application

               -Large networks

               -Simple administration

               -Internet/WAN access

Single-Server Network (10-50 users):

• If you have fewer than about 50 people, you can run your entire organization with a single server.
• This allows you to centralize a number of services and maintain strong control over your network environment.

Single server nets are good for:

-Centralized file services

-Network printing

-E-mail

-Work flow and groupware

-Login security

-Archiving

-Organizing data

-Easy installation

-Simple administration

They are not good for:

               -Application serving

               -Distributed organizations

               -Large organization

Multiserver Networks (50-250 Users):

As your network grows, however, a point will come when you need to begin adding more servers.

Multiserver nets are good for:

-Centralized file services

-Networking printing

-E-mail

-Workflow and groupware

-Login security

-Application services

-Large database

-Internet/WAN access

They are not good for:

               -Tight budgets

               -Easy installation

               -Organizing data

               -Simple administration

Multiserver High Speed Backbone Networks (250-1000 Users):

• With more than 250 clients, network planning becomes a lot more challenging.
• This number of clients tends to be spread out over larger areas than can be supported from a central computer room.
•  This geographic aspect requires both a distributed network and a lot of servers.
• A network of this size will be connected with a high speed backbone that runs between servers.

It is good for:

-Centralized file services

-Networking printing

-E-mail

-Workflow and groupware

-Login security

-Application services

-Client-server database

-Internet/WAN access

It is not good for:

               -Tight budgets

               -Easy installation

               -Organizing data

               -Speed

Enterprise Networks (1000+ Users):

• Enterprise networks are so large they are no longer really considered a single network.
• With more than 1000 users, it’s best to break down the network into multiple connected networks that have different directory services and are split along some natural boundary.
• These smaller networks are then designed according to the criteria presented above smaller networks and then connected with the network and internetwork connectivity services.

It is good for:

-Networking printing

-E-mail

-Workflow and groupware

-Login security

-Application services

-Client-server database

-Internet access

It is not good for:

               -Tight budgets

               -Easy installation

               -Centralized file services

               -Organizing data

               -Speed

How far Apart Are the Computers?

• The distance to the most distant client computer is important; it will help you determine which network protocol you should use and what type of cabling will work for your situation.
• Walking Rule: If you are less than 6 ft  tall, multiply the number of paces by 2 ft. If you are over 6 ft, multiply the number of paces by 30 inches; then divide by 12 to get the number of feet. Add 20 ft for vertical rise to the ceiling and back down. The result will be a good estimate of the number of cable-feet between your most distant client and your network equipment area.

What software are you using?    

• Software and files are the data that flows over a network, so knowing what type of software is in use will give you a good estimate of how much data per client will traverse the network.
• Word processors and Spreadsheets, Graphics and CAD, Database software

What software will you use?    

• An important software consideration in addition to the software currently in use is the software you will add once your network is up and running.
• Networks naturally improve the communications processes in networked organizations with tools such as e-mail, internet connectivity and groupware.

What Special Requirement do you have?       

• Do you have some special need for security on your network?
• Are any of your computers more than 100 m from where you will locate your hub?
• Have there been any problems with electrical interference that you know about?
• These issues will play a part in determing what sort of cabling and network devices you will need to install.

How much can you spend?

• The amount of money you can spend is a factor that will determine which solutions are available to you.
• Be sure to take the time to determine how much money you will be able to justify spending on your network.

Data Segmentation and Token Ring For Networking.

Segmentation:

• As an Ethernet network grows and more stations are added to the LAN, performance can drop significantly.
• Ethernet is a shared media network; when a lot of stations have data to transmit, the net gets congested and many collision occur.
• Segmentation is the solution.
• Segmentation is the process of splitting a larger Ethernet network into two or more segments linked by brides or routers.
• The resulting segments have fewer stations to contend with for access to the net and the router or bride transfer data from one segment to the other only when the destination for the data is on the other segment. The rest of the net traffic stays in the segment where it belongs.

Token Ring:

Token Ring was developed by IBM as a robust and highly reliable network.

Specification:

Cable type                                                                  UTP, STP or fiber-optic

Maximum MSAUs                                                      33

Maximum nodes                                                         260

Max dist between node and MSAU                            45.5 m for UTP, 100m

Max patch cable distance connecting MSAUs             45.5m UTP, 200m STP, 1 Km fiber-optic

Min patch cable dist connecting MSAUs                      2.5m

Max cumulative patch cable dist connecting all MSAUs   121.2m UTP, several Km for fiber-optic

How Token Ring Works:

• The ring passes a free token around the ring in one consistent direction.
• A node receives the token from its nearest active upstream neighbor and passes it to its nearest active downstream neighbor.
• If a station receives a free token, it knows it can attach data and send it on down the ring.
• Each station is given an equal chance to have the token and take control in order to pass data.
• Each station in the ring receives the data from the busy token with data attached and repeats the token
• and data, exactly as it received them, to the next active downstream neighbor on the ring.
• The data is received and retransmitted by each node on the network until it has gone full circle.

Advantages and Disadvantages:

• Token Rings continues to operate reliably under heavy loads
•  Built-in diagnostic and recovery mechanisms
• Token Ring makes connecting a LAN to an IBM mainframe easier
• Fault-tolerance features are provided through ring configuration

Disadvantages:

• Token Ring cards and equipments are more expensive than Ethernet
• Token Ring can be very difficult to troubleshoot and requires considerable expertise

FDDI

FDDI:

• Fiber Distributed Data Interface (FDDI) is another ring-based network.
• FDDI uses fiber-optic cables to implement very fast, reliable networks.
• FDDI uses a token-passing scheme to control network access.
• Several FDDI devices can transmit data simultaneously.
• A token is passed around the ring, and the possessor of the token is allowed to transmit FDDI frames.
•  A FDDI network may have several frames simultaneously circulating on the network.
• This is possible bcz the possessor of the token may send multiple frames, without waiting for the first frame to circulate all the way around the ring before sending the next frame.
• The possessor of the FDDI token is also allowed to release the token and send it to the next station in the ring as soon as it is through transmitting frames, rather than having to wait for the frames to make it all the way around the ring.