Unit 10 Networks.

Unit 10 Networks

The basic components of a network:
Connections (Physical Medium) Protocols (Common language) Services (Clients, Servers, and Files) In its most basic form, a network consists of three items. These are: The connections, or the physical medium, Protocols, which form a common language, and services which include the role of a client, a server, and files. Let’s take a look at each of these components.

The Client is the computer requesting the use of a resource.
May I have Worksheet-101? The client is the computer or workstation that requests the use of a resource or service. Here the computer on the left is requesting a worksheet. This makes the computer on the left the client. Client Server

The Server is the computer providing the resource.
Sure, here it is! Server The computer on the right responds by sending the client the requested worksheet. The computer that provides the resource is called the server. Here, the computer on the right is the server because it “serves up” the resource requested by the client. Client Worksheet-101

In some networks, every computer acts as both Client and Server.
In some networks all workstations are equal. They act as both clients and servers.

Peer-to-Peer Network Client & Server Client & Client & Server Server
A network in which each workstation acts as both a client and a server is called a peer-to-peer network. In the peer-to-peer network all computers share and share alike. The responsibility for control and security is distributed throughout the network.

Peer-to-Peer Network No centralized data access.
Security handled by assigning each resource a password. There is no centralized data access or security control. Everyone is more or less equal and there is a kind of mutual trust among the users. Security is achieved by assigning each resource a password… a password that is revealed only to those users who need to share that particular resource.

Peer-to-peer works best with a a small number of computers.
Obviously, if you have many different users, each with various resources to be shared, there is a blizzard of passwords. For this reason, peer-to-peer networks are usually quite small, typically having ten workstations or less. As long as they are small and security is not a major issue, peer-to-peer networks are an inexpensive, easy-to-administer solution to resource sharing.

One computer can act as server for several clients.
For larger networks, or in situations where security is paramount, a better arrangement is a centralized server that handles data access and security for several clients. Such an arrangement is called a Client/Server network. Here a single server controls a number of clients.

Client/Server Network
Centralized data access. Centralized security. Centralized administration, record keeping, and control. The client/server network is used in situations where centralized control of critical data is important, such as in businesses, government, hospitals, etc.

Larger networks require the client/server arrangement.
Because of the shortcomings of the peer-to-peer network, larger networks require the client/server arrangement.

The Physical Medium is the channel the data travels between computers.
Client Now, returning to our list of basic network components, we mentioned connections, or the physical medium. Let’s consider the physical medium. The physical medium is the channel that the data travels between computers. Server Physical Medium

The Physical Media can be:
Wire Fiber Optic Cable Wireless Link (RF or Infrared) Hardware devices such as network cards, hubs, routers, etc. The physical media includes the wire, cables, and electronic devices that transport the data from one computer to another. It may consist of copper wire, fiber optic cables, or even wireless links such as radio waves or infrared. It also includes the hardware devices that facilitate data transfer such as network interface cards, modems, hubs, routers, switches, etc. We will discuss these in detail in future lessons.

The common language is the Network Protocol.
May I have Worksheet-101? Sure, here it is! Client The final of our basic components is a common language used by the computers in the network. This common language is called the network protocol. Here, we see the Client asking for a worksheet and the server replying with the requested data. The protocol determines everything from the nature of the request, to the speed of the reply, right down to such basics as what constitutes a logical 1 or 0. Server

Protocol A signed document containing the record of the points on which agreement has been reached by negotiating parties. The code of ceremonial forms and courtesies accepted as proper and correct in official dealings. Webster’s New World Dictionary defines protocol as: A signed document containing the record of the points on which agreement has been reached by negotiating parties. While this definition was not written with computers in mind, it is a reasonable definition for Networking Protocols as well. The Dictionary goes on to define protocol as: The code of ceremonial forms and courtesies accepted as proper and correct in official dealings. Once again, not a bad definition for Networking Protocol.

In Networking, Protocols are:
Agreements that describe how things work. Industry-wide frameworks that describe every aspect of communications between computers. Grouped together into Protocol Suites. In networking, protocols are agreements about how things work. They are industry-wide frameworks that describe every aspect of communications between computers. Networks rely on protocols to communicate in much the same way that humans rely on language. There are so many different protocols that they are often grouped together into Protocol Suites. A Protocol Suite is a collection of protocols that work together. For example, the TCP/IP Protocol Suite contains over 100 different protocols. One of the most basic decisions in planning and building a network is choosing the Protocol Suite.

The main Protocol Suites:
NetBEUI – Small Microsoft networks. SPX/IPX – Primarily Novell networks. TCP/IP – Must be used if the network is to connect to the Internet. In the PC world, the three main Protocol Suites are: NetBEUI, which is used in small Microsoft-based networks (but not widely used by Windows XP). SPX/IPX, which is used primarily in Novell-based networks. TCP/IP, which must be used if the network is to connect to the Internet. But it is also widely used even in those networks that do not connect to the Internet. Today many networks have evolved from smaller departmental networks or through mergers and acquisitions of whole companies. For this reason, many networks do not use just one Protocol Suite, but rather all three of those listed here, along with others, such as Apple’s LocalTalk if Macintosh computers are used in the network.

One of the most basic ways to classify networks is by their topology.
Earlier, we saw that networks may be classified as peer-to-peer or client/server depending on whether control is distributed between all workstations or is centralized. Another basic way to classify networks is by their topology.

Topology refers to the way the computers in the network are connected to each other.

The most common topologies are:
Bus Star Ring Mesh You should be aware of four different topologies. These are the bus, the star, the ring, and the mesh topology. Let’s take a look at each of these.

Bus Topology “T” “Backbone”
Here we see an example of the bus topology. Notice that all the workstations are connected to a single cable, sometimes called a backbone. Notice also that each workstation connects to the cable via a “T.”

Bus Topology A single cable interconnects all workstations.
The cable is terminated at both ends. Terminator Terminator You can recognize the bus topology by the single cable that interconnects all the workstations. Another characteristic is that both ends of the cable must be terminated by a special device called, appropriately enough, a terminator. We will discuss all this in much greater detail later.

Star Topology The star topology uses a different connection scheme. Here the workstations are connected to a central point like the spokes in a wheel or the points of a star.

Star Topology All workstations connect to a single central hub. Hub
This is its most distinguishing characteristic and an easy way to recognize the star topology. The central point is actually a device such as a hub or a switch.

Ring Topology Yet another connection scheme is the ring topology. Here each workstation is connected to the computer before it and the one immediately after it. The most common implementation of this topology is called token ring. An electronic token passes from one computer to another around the ring. When a workstation has the token, it is that computer’s turn to transmit data. This arrangement prevents two computers from attempting to use the media at the same time.

Ring Topology Backbone is a ring.
Each workstation connects only to two other workstations. To identify this topology, look for the ring connection. There should be no end to the backbone. Notice also that each workstation connects to only two computers, the one in front of it and the one in back.

Mesh Topology A final connection scheme is the mesh topology. Here there are multiple paths between workstations. In the simplistic view shown here, you can imagine each computer having multiple network interface cards with multiple cables leaving it. And while such an implementation is possible, it becomes very expensive and unbelievably complex if more than a handful of workstations participate.

Mesh Topology Net-1 Net-2 Router Net-3 Net-4
Generally, the multiple paths occur at a higher level in the network. Devices called routers are used to route messages between networks. Often routers are connected together in such a way that multiple paths exist from one network to another.

The Internet is an example of a Mesh topology.
The Internet is the most striking example of mesh topology. Once inside the public telephone system, shown here as the cloud, there are multiple paths to each workstation.

Another way of classifying networks is by their size or geographical span.
Yet another way to classify networks is by their size or geographical span. Some networks are confined to a single room. Others span the entire globe. So a geographical span classification helps us to visualize the size of the network.

The most common size classifications are the:
Local Area Network (LAN) Metropolitan Area Network (MAN) Wide Area Network (WAN) The most common size classifications are shown here. They include the Local Area Network or LAN, the Metropolitan Area Network or MAN, and the Wide Area Network or WAN.

The Local Area Network (LAN)
As the name implies, the LAN is confined to a local area, usually a single building.

Metropolitan Area Network (MAN)
On the other hand, the MAN may connect several different networks over a large metropolitan area.

Wide Area Network WAN Finally, a wide area network connects networks over a very large geographical area such as a state, an entire country, or even the whole world.

Networking Components

Components Network Operating System Interface Cards Cables
The three networking components that you are most likely to use are the Network Operating System, the Network Interface card, and the cables. Let’s look at each of these in some detail, starting with the operating systems.

Network Operating System (NOS) vs. desktop Operating System (OS)
The Network Operating System is to the desktop operating system as the lion is to the house cat. They are both members of the same family, but one is vastly more powerful and more likely to survive in the wilderness. Let’s take a look at some of those benefits again and see how the Network Operating System (NOS) is superior to the typical desktop Operating System (OS).

NOS provides improved security
User-level Security Server authenticates: User name User password User location Server determines which resources the user may access. Simplifies security for the user. Most Network Operating Systems offer improved security in several ways. First, user-level security is used instead of share-level security. With user-level security, the user must log on to the NOS before she can use the network. The NOS then authenticates that: the user is authorized to use the network; she knows the right password; and she is at a permissible location. The NOS also keeps strict track of which resources the user is allowed to access. This centralized system is much more secure than the distributed method used with peer-to-peer networks. While the NOS increases security many-fold, it makes security procedures more convenient for the user. In most cases, the NOS allows the user to get by with a single password. And, the NOS allows the user to select that password.

The NOS allows sensitive resources to be maintained in a centralized/secure location.
The NOS also allows you to keep sensitive resources in a centralized and secure location. It allow you to put all your eggs in one basket and then lock that basket away for safekeeping. Having important data in one location also makes it much easier to back up that data.

The NOS provides improved performance.
Client is relieved of the burden of server Server can be optimized Adds administration and management capabilities. A true Network Operating System improves performance in several ways. First, it allows client machines to perform better because they are once again single-user machines which are no longer burdened by serving other clients. Second, the NOS allows a very powerful machine to be selected as the server. The NOS may support multiple processors, special arrays of hard drives, special tape backup systems, uninterruptible power supplies, etc. Finally, the Network Operating System (NOS) generally has many additional management and administrative capabilities not found in a desktop OS.

The NOS provides better administration.
Centralized security Centralized data Consistent policies Administrative tools Accountability The NOS offers much better administration than does its desktop OS counterpart. Security policies can be administered uniformly and consistently from a centralized location. Keeping the data in one place and under the control of a central server makes it easier to protect and to back up. The special administrative tools provided by the Network Operating System make it easier to keep track of users, monitor security, and spot overloaded resources and bottlenecks. Finally, various reports of the users’ and the administrator’s activities provide accountability. In addition, because of the centralized and complex nature of the NOS, someone is assigned the task of Network Administrator. He or she usually becomes the primary point of accountability for the network.

The NOS allows scalability
The client/server network grows gracefully. A server may handle over 50 clients. Additional servers can be added. Specialized servers File server Print server Communications Server The NOS offers excellent scalability. It allows a single server to handle dozens of clients. And it allows additional servers to be added as needed. Moreover, it allows a server to be optimized to perform a particular task such as file server, print server, or communications server.

Disadvantages of the NOS
More expensive More complex Requires administration. All of these advantages come at a price. The typical network operating system is more expensive, more complex, and it usually requires someone to act as an administrator, if it is to be used to full advantage. These are the prices you must pay for a large, robust, and secure network.

Major Network Operating Systems
Unix Novell NetWare Windows NT Server The three major network operating systems are Unix, Novell’s NetWare, and Microsoft’s Windows NT Server. Let’s take a brief look at each of these.

Unix Multitasking, Multi-user OS
Influential in evolution of the Internet Uses TCP/IP Clients communicate with Server using Terminal Emulation Unix was one of the first multitasking, multiuser operating systems. A mainstay of university and scientific computing, it has set the standard for server-based network operating systems. It was particularly influential in the evolution of the Internet. In fact, the Internet originated as a handful of Unix servers at universities scattered around the globe. Unix pioneered the TCP/IP protocol, which has become the protocol of the Internet. Originally, dumb terminals were used to communicate with the Unix operating system. They had just enough intelligence to operate a video display, keyboard, and the network interface. All of the applications resided on the server. Today’s Unix-based networks often use computers that are configured to emulate the operation of a standard terminal. These client computers use the TCP/IP protocol.

Linux Smaller, easier to use version of Unix
Easily adapted to individual computers or as server in a small network Uses the command line as the user interface. Offered with a graphical user interface. Because of its adaptability and universal acceptance in the computer industry, several different universities and corporations have developed versions of Unix that operate in a similar fashion. However, most are expensive and are designed primarily for large computer networks. For those who don’t need the power of a full-blown Unix, there is a smaller, easy to use, and virtually free version called Linux. Offering the same basic features of Unix, Linux is easily adapted to both individual computers as well as small networks. While both Unix and Linux are primarily command-line driven, several organizations are offering a Windows-type interface for Linux.

Novell’s NetWare Text-based operating system
Supports all Windows-based client computer operating systems IPX/SPX is the primary NetWare protocol Another very popular Network Operating System is NetWare from Novell Corporation. Novell’s NetWare is used in more businesses than any other NOS. Until the introduction of Version 5, NetWare was a text-based operating system. That is, issuing a command or changing a function was handled in one of two different ways. You could use command-line entries like you would in Unix, or you could open DOS-like menus to perform the operation. NetWare 5 introduced a Windows-like interface. At the time of this writing, most NetWare installations continue to use different levels of NetWare 3 and 4. NetWare supports several different Client computer operating systems, virtually all of them are Windows-based. Both NetWare 3 and 4 require the client computer to support the IPX/SPX protocol. Beginning with NetWare 5, the client computer may use the TCP/IP protocol.

Windows NT 32-bit network operating system
First to use a graphical user interface Not as feature-rich as Unix and NetWare Supports a wide range of network clients The third major Network Operating System is Windows NT. Windows NT is a 32-bit network operating system. It was the first to use a graphical user’s interface. It doesn’t provide all of the features of the other network operating systems. But it is becoming a favorite NOS for small networks because of its price and its compatibility with the other operating systems. By compatibility, we mean that Windows NT does an excellent job of supporting all of the network clients. These include all of the Windows, DOS, OS/2, Macintosh, NetWare, and Unix clients. Like all server software, Windows NT provides network resource management and control.

The Network Interface Card is also called:
The Network Card The Network Adapter The Network Adapter Card The NIC The Network Interface Card has several names. Some call it simply, the network card. Others call it a network adapter; still others, the network adaptor card. In this course, we will generally use the name network interface card or more likely, the abbreviation N-I-C, or by actually pronouncing the abbreviation as “nick”. Whatever, we call it…

The NIC is the main interface between the computer and the network cable.

Parallel data into serial data.
CPU Memory int e l Computer Recall that most internal operations within the computer involve parallel data. That is, data is moved in blocks of 8, 16, or 32 bits at a time. However, the network cable usually has a single conductor for carrying data. Therefore, serial data must be placed on the cable one bit at a time. So before data can be transmitted over the network, the data must be converted from parallel to serial form. This is one of the key functions of the NIC. Cable NIC

Considerations when selecting the NIC:
Type of cable connection Type of Address/Data Bus The method used to configure the card. There are three important considerations when selecting the NIC: What type of network cables are used? Obviously the cable connector on the card must match that of the cables used in the network. Also of importance is: Where does the card plug into the motherboard? Both ISA and PCI network interface cards are common. Finally, as you will see, there are several different ways of configuring the NIC. Let’s look at each of these in more detail.

The NIC must match the cable.
RJ-45 This is what a typical Ethernet NIC looks like. The card shown here can accommodate two different types of network cables. The BNC connector accepts a coax cable, while the RJ-45 jack accepts a twisted-pair cable. You use one type of cable or the other, not both. Some NICs will accept one type of cable only. In any event, make certain that the card matches the cable. BNC

RJ-45 Connector The RJ-45 connector is the connector of choice for twisted-pair Ethernet cables. Here you can see one attached to an unshielded twisted-pair cable. The transparent connector lets you look inside. It is similar to the connector on an ordinary telephone, only larger, and with eight wires instead of four. We will examine it in more detail in a future lesson.

The NIC must match the Address/Data Bus of the computer.
ISA Another way to identify a network interface card is by its address/data bus. There are several possible types, including ISA, EISA, MCA, VL-bus, and PCI. However, if you examine the computers manufactured over the past couple of years, you will find only two busses: the ISA bus (shown above) and the PCI bus (shown below). (Of course, new computers include a video processor AGP bus, but we’re talking networks, not video.) In fact, the latest computer standard from Microsoft eliminates the ISA bus from the motherboard, leaving only the PCI bus for everything except video. If you ever service a computer with an EISA, MCA, or VL-bus interface, you may be in for a shock. It is virtually impossible to find NICs with these types of busses. In any event the NIC must match the bus you are going to plug it into. PCI

Installing the NIC in the computer.
NIC installed inside the computer. Normally plugs into a bus slot. Some are built right into the motherboard. Plug-in boards must be configured correctly. The NIC installs inside the computer, usually plugging into a slot in the motherboard. Some computers even come with the NIC components installed directly on the motherboard. Often the NIC is the only physical part of the network that is inside the computer. Everything else is usually external. And like other plug-in boards, the NIC must be configured correctly.

NIC Configuration Methods
Plug-and-Play EEPROM Jumper pins Which brings us to a third consideration in selecting the NIC. There are three basic ways to configure the NIC. If the NIC and the computer in which you are installing the NIC support Plug-and-Play, then you have nothing to do. The computer takes care of the configuration for you. On the other hand, a NIC that uses EEPROM (Electrically Erasable Programmable Read Only Memory) to store the configuration information must be programmed by you through a software utility. That means you have to know what resources to assign to the NIC. The third method for configuring a NIC uses jumper pins to assign computer resources.

Preparing to Install the NIC
Ensure there is an open bus slot. Ensure the adapter is compatible. Ensure there are system resources available. Ensure all installation items are available. Ensure all software is available. Now let’s talk about how to install a NIC in a typical personal computer. It’s pretty much like installing any other adapter. Here’s a quick review. Beginning with the obvious, make sure there is an open bus slot. Computers with both ISA and PCI bus slots almost never accommodate as many adapters as there are slots. That’s because one ISA and one PCI bus slot share one rear panel opening. You’ll always wind up with one unusable bus slot. You must also make sure the adapter you’ve selected is compatible with the open bus slot, the network cables, and the network technology you’ll be using. Next, check on the system resources. With today’s multifunction computers, it’s easy to use up the available IRQs or DMA channels. Then again, you may have available resources, but the adapter can’t be configured to use those particular resources. In addition to the adapter, you’ll also need several other items. These may include external transceivers, a T-connector, cable, hub, and system documentation. Finally, you need all the software to get the network interface up and running. This can include the adapter driver and support utilities, network shell, and diagnostics. Most of this software is supplied with the operating system. Adapter-specific software usually comes with the adapter.

Installing the NIC Hardware
Configure the NIC to available resources. Use a ground strap. Remove cover from the computer. Remove rear panel slot cover plate. Remove card from its antistatic bag and immediately plug it into motherboard. Secure card slot cover plate to computer. You’ve installed cards before, so these steps are just reminders. The first step is valid only if you are dealing with a NIC that uses jumpers or dip switches to configure the resource allocation. The next step should be pretty much automatic. Static electricity can be your worst enemy when working with exposed computer circuits, so use the grounding strap. Now you can safely remove the computer cover and the rear-panel, slot-cover plate. When you remove the NIC from its antistatic bag, try to hold it by its edges. Having taken the NIC out of the antistatic bag, don’t set it down. Instead, immediately plug it into the motherboard card slot. Make sure the card is fully seated in the connector, then secure the card to the rear panel of the computer.

Installing the NIC Software
Loading the device driver used by the NIC. Loading any utilities supplied with the NIC. After the card is physically installed, its software programs must also be loaded. The first is the card’s device driver which allows the card to communicate with the operating system. Second are the utilities that are usually supplied with the card. This usually includes diagnostic software that is unique to the NIC and a setup program for cards that use EEPROMs for configuration purposes.

NIC Device Driver Supports communication between the NIC and OS.
Automatically installed and configured if both NIC and OS support PnP. In other cases, driver loaded from floppy or CD supplied with NIC. The NIC device driver allows the NIC and the computer operating system to talk to each other. Depending on the operating system and the type of NIC involved, there are different methods of installation. With plug and play, everything may happen automatically. Operating systems like Windows 98 or Windows NT come already equipped with many common NIC drivers. If your NIC has a driver already supplied in the operating system, you will have little to do. On the other hand, in non-plug-and-play situations, you may have to install the device driver from a floppy disk or a CD-ROM supplied with the NIC.

A more recent device driver may be available at the website of the NIC manufacturer.
Either way, the device driver may not be the most recent one available. To get the most recent device driver, go to the website of the NIC manufacturer.

Troubleshooting the NIC
Is NIC talking to the motherboard? Is the NIC working internally? Is the NIC communicating with the external network? All too often, after installing a new NIC, the new card refuses to connect to the network. When this happens, you need a systematic approach to troubleshooting the problem. One such approach is to ask these three key questions. First: Is the NIC talking to the motherboard? Here you need to check for resource conflicts. In Windows 98, the Network Adapter Systems Properties dialog box in the Control Panel has a tab called Resources that will tell you the IRQ and I/O Address used by the NIC, and can alert you to resource conflicts. The NIC usually comes with diagnostic software to answer the last two questions: Is the NIC working internally? And: Is the NIC communicating with the external network?

LEDs Link Other useful indicators are the LEDs on the NIC. Different cards provide different numbers of LEDs to display card status. This one has two LEDs. The top one lights to show there is a link between the NIC and a node at the other end of the interface cable. This does not tell you the NIC is talking to its host PC, so don’t make that assumption when troubleshooting a network problem. The second LED on this NIC shows communication activity. Again, you can’t make too many assumptions about this LED. When it lights, it’s telling you data is passing through the cable attached to the NIC. It doesn’t tell you if the data is entering or leaving the NIC. Activity

Twisted-Pair Cable UTP—Unshielded Twisted-Pair
STP—Shielded Twisted-Pair Next, let’s discuss the cables that connect all the network components together. There are two types of twisted-pair cable: Unshielded Twisted-Pair (UTP) and Shielded Twisted-Pair (STP). UTP cable is used for more purposes than virtually any other type of cable, from telephones to high-speed networks. UTP cable usually contains multiple pairs of twisted wires. Bundles of 2, 4, 6, 8, 25, 50, and 100 twisted-pairs of wires are common. Like UTP cable, Shielded Twisted-Pair (STP) cable is made up of two or more twisted-pairs of wires. However, in STP cable, the wires are electrically shielded, much like the center conductor of a coaxial cable.

Twisted-Pair Cable STP UTP Foil Shield Wire Braid
Here are some examples of UTP and STP cables. The STP cable on the left has a foil shield that wraps around all four of its twisted-pairs of wires. The UTP cable in the middle is identical to the STP cable on the left, except for the lack of shielding. The STP cable on the right has only two pairs of twisted wires. But each pair of wires is wrapped in its own foil shield. Thus, each twisted-pair is electrically shielded from the other. Then everything is wrapped in a wire braid to strengthen the cable and complete the shielding process. Because of the extra material involved and the extra manufacturing steps required, shielded twisted-pair is inherently more expensive than unshielded twisted-pair. Also, the extra shielding material makes the cable larger, stiffer, and harder to install.

10BaseT Ethernet uses Unshielded Twisted Pair (UTP) cable.
For this reason, 10BaseT Ethernet uses Unshielded Twisted Pair (UTP) cables. The cable contains four twisted-pairs, for a total of eight wires.

RJ-45 Connector Strain Relief Crimp Contacts Wires Latch
The RJ-45 connector is the connector of choice for twisted-pair Ethernet cables. Here you can see one already attached to a UTP cable. The transparent connector lets you look inside. It is very similar in appearance to the connector used with telephones, only larger, and with eight wires instead of four. Wires Latch

Twisted-Pair Advantages
UTP is inexpensive. Workstations isolated from each other by central hub. Easy to add workstations to segment. Easier to route than coaxial cable. Easier to troubleshoot than coaxial cable network. Twisted-pair cable offers several advantages compared to its nearest competitor, coaxial cable. First of all, it is cheaper to manufacture. Second, it’s not as expensive to install. Recall that workstations are isolated from each other by a central hub. For that reason, it is easy to install additional workstations to a local network segment without adding to the overall length of the cable segment. Because of the way twisted-pair is constructed, it’s a lot easier to route than coaxial cable—especially around corners. Finally, twisted-pair is easier to troubleshoot than coax because each workstation is isolated from the others by the hub.

Twisted-Pair Disadvantages
UTP cable is susceptible to RFI/EMI. Suffers crosstalk between wire pairs. Poor conductor; attenuates signal more quickly than coax. Maximum segment length half of coax. STP cable more expensive than UTP cable or coax. Difficult to work with shielding. Even so, twisted-pair cable is by no means perfect. It has several disadvantages. While STP cable offers a high degree of shielding from outside electrical interference, UTP cable has no shielding, and therefore, is very susceptible to electrical noise. In addition, both UTP cable and STP cable suffer from cross-talk between wire pairs. The amount of cross-talk is affected by the type of cable and the frequency of the signals. Higher category cables and lower frequencies are less affected by cross-talk. Compared to coaxial cable, twisted-pair cable is not a good conductor. It attenuates the signal more quickly than coax. In fact, the maximum segment cable length before the signal must be regenerated is only 100 meters, approximately half that of thinnet cable. From an economic standpoint, UTP cable is the least expensive, thinnet cable is next, and STP cable is most expensive. Likewise, of the three types of cable, STP cable is the most difficult to work with. That’s because of its shielding and the resulting stiffness.

EIA/TIA Cable Categories
Category 1—Voice-grade UTP phone Category 2—Data-grade UTP, 4 Mbps Category 3—Data-grade UTP, 10 Mbps Category 4—Data-grade UTP, 16 Mbps Category 5—Data-grade UTP, 100 Mbps In most applications, the benefits of UTP far outweigh its disadvantages, and 8-wire UTP is used in virtually all 10BaseT installations. There are many types of UTP cable. However, the cables you need to know about are part of the collection specified by the Electronic Industries Association and the Telecommunications Industries Association (EIA/TIA). They are called Category 3, or Cat 3, cable and Category 5, or Cat 5, cable.

Category 3 Cable Common data-grade cable.
Four unshielded twisted-pair wires. Transmission rates up to 10 Mbps. EIA/TIA Category 3, or Cat 3, cable is a data-grade UTP cable that was used in a lot of older computer network installations. The term data-grade means that each twisted-pair of wires in the cable can carry a digital signal without distorting the signal to the point that it will not be usable. Voice-grade cable can only support analog signals and extremely slow (less than a megabit-per-second) digital signals. There are four unshielded twisted-pair wires in a Cat 3 cable. The cable design—the thickness of the copper conductors, the type of insulation, and the number of twists per foot of twisted-pair wires, and how the twisted-pair wires are arranged inside the cable jacket—determine the maximum data transmission rate of the cable. In the case of Cat 3 cable, that rate is up to 10 megabits per second.

Category 5 Cable Data-grade cable.
Official transmission rates up to 100 Mbps. Four unshielded twisted-pair wires. EIA/TIA Category 5, or Cat 5, cable is also a data-grade UTP cable. In fact, Cat 5 (or better) should be the only UTP cable that is installed today. While its initial cost may be a few cents more per foot, its installation labor cost is the same and it allows for future upgrades to higher speed networks. In addition to new installations, many older Cat 3 cable networks are switching over to Cat 5 cable. That’s because Cat 5 cable can support data transmission rates up to 100 megabits per second. Like Cat 3 cable, Cat 5 contains four unshielded twisted-pair wires.

Two Bus Technologies: 10Base5 or Thicknet 10Base2 or Thinnet
Next, let’s look at coaxial cables, which are normally used as a “bus” or “backbone” in a network. There are two different flavors of the bus topology that we will discuss in this presentation. They have the rather strange titles of 10Base5 (Pronounced “Ten Base Five”) and 10Base2 (Pronounced “Ten Base Two”). 10Base5 is the original implementation of Ethernet. It is covered for historical purposes and because some of the old 10Base5 LANs are still in use. Generally though, when the bus topology is used today, it is a 10Base2 implementation rather than the older 10Base5. Because of the types of cables used, 10Base5 is often called Thicknet, while 10Base2 is often called Thinnet. Let’s begin by taking a closer look at these cables.

Coaxial Cable Characteristics
RG-58 Thinnet Cable Sleeve Wire Braid Dielectric Center Conductor The bus topology uses coaxial cables like those shown here. The one on top is RG-58 coax, also known as thinnet cable. The one on the bottom is RG-8 coax, also known as thicknet cable. The thick and thin names refer to the relative sizes of the cables. The coaxial cable has a single center conductor that is surrounded by an insulating material called a dielectric. The dielectric is, in turn, surrounded by a layer of foil or wire braid. This is then covered by an outside insulating sleeve that’s often called a jacket. Network data passes through the center conductor in the coaxial cable. The foil or wire braid is grounded and serves two purposes. First, it provides a signal return path to complete the circuit. Second, it shields the cable from external electrical interference. Foil RG-8 Thicknet Cable

Coax Advantages Resistant to RFI. Good conductor.
Longer network segments. Thinnet: easy/inexpensive workstation interconnection. Thicknet: stronger, more durable than any other network cable. Coaxial cable offers several advantages compared to its nearest competitor, twisted-pair cable. First of all, it is very resistant to outside electrical interference. That’s because of its wire braid or foil shield. Next, it is a better conductor over long distances because there is less resistance in the signal conductors. This allows us to build longer network segments using coaxial cable. In the case of the thinnet networks, they are easier and less expensive to build. Mainly that’s because the cable is directly connected from one workstation to the next, rather than from each workstation to a central hub, as is the case with a twisted-pair cable network. The heavy braid in the thicknet cable makes the cable physically stronger than any other network cable.

Coax Disadvantages Each segment must be terminated.
Break in cable disables entire segment. Difficult to add workstations to a thinnet segment. Thicknet: difficult to use, heavy shield braid and foil. Thicknet: requires special piercing tap and transceiver. Even so, coaxial cable is far from perfect. It has several disadvantages. Each end of the cable must be terminated with a resistive load to prevent signal reflections. Because of the need for proper termination, a break in a cable segment disables all of the workstations in that segment. While it is easy to build a local thinnet network, adding workstations to that local segment can be very difficult because of the way the cable is routed from workstation to workstation. Breaking the segment to add a new workstation temporarily disables the entire segment. Moreover, the new cable may add excessive cable length to the segment, causing the signal to degrade. Thicknet cable has even greater problems. For starters, it is difficult to work with. It doesn’t bend easily because of its size and the heavy shielding inside. As a rule, thicknet is used as a backbone to interconnect several local network segments. Access to the thicknet signal is gained by attaching a piercing tap that bores a hole into the cable and makes separate connections with the center conductor and the shield. A special transceiver completes the connection to the local network segment. The piercing tap is often called a vampire tap. Another name for the transceiver is a media converter.

What’s in a name? 10Base5 10Base2
Thicknet and Thinnet have other names, as we have seen. Thicknet is called 10Base5, while Thinnet is called 10Base2. What’s in a name? Actually, quite a lot in this case. Each name has three parts and each part has its own meaning.

The number on the left is the speed of the LAN.
10 Base 5 10 Base 2 Speed in Megabits Per Second The number on the left stands for the speed of the Ethernet LAN in megabits per second. Notice that the speed is expressed in megabits per second, not megabytes per second. Both the 10Base5 and 10Base2 LANs have a maximum speed of 10 megabits per second.

The number on the right is the length of the LAN segment.
10 Base 5 10 Base 2 Length of Segment in Hundreds Of Meters The number on the right is the maximum length of a single segment of the LAN in hundreds of meters. Thus the maximum length of a 10Base5 segment is 500 meters. Unfortunately, in the case of 10Base2, the naming convention fails slightly. It turns out that the maximum segment length is not 200 meters, as one might expect from the 2, but rather 185 meters. Why the discrepancy? Perhaps the number was rounded to avoid the name 10Base1.85. In any event, it is important to remember that …

The word in the middle signifies the type of signal.
10 Base 5 10 Base 2 Type of Signal But to get back to the naming convention, we still have the middle term. In this case: Base. The middle term refers to the type of signal used. Base stands for baseband signal. This is a fancy way of saying that a single signal is used.

Baseband Broadband Signal 1 Signal 2 Signal 3
Here we see a comparison of baseband versus broadband. Once again, baseband means that a single signal is transmitted on the cable, as shown at the top. Another situation called broadband is possible. Here a number of different signals at different frequencies are transmitted simultaneously on the cable. The most common application of broadband is cable television, where 60 or more channels may be transmitted on a single cable. At present broadband techniques are not widely used in networking, although they may become more common as cable TV companies add networking capabilities to their systems. Signal 1 Broadband Signal 2 Signal 3

RG-8, 10Base5, or Thicknet Cable
When the bus topology is implemented today, it will most likely use 10Base2 technology. As you have seen, the 10Base2 uses the RG-58 cable. Compared to the RG-8 Thicknet cable, the RG-58 cable is quite thin and flexible. This makes it much cheaper and easier to work with than Thicknet. The name thinnet comes from the RG-58’s trimmer size. RG-58, 10Base2, or Thinnet Cable

The RG-58’s lighter shield still provides good protection against electrical noise.
While the RG-58 cable has a lighter shield than the heavier RG-8, its shielding is still adequate for most applications. In fact, it provides better shielding than the more popular 10BaseT technology which we will discuss in future lessons.

Fiber Optic Cables Finally, let’s look at the newest type, fiber optic cable. Fiber optic cable is the best cable type, in terms of loss, cable lengths, resistance to electrical interference, and is impressive data transmission rates. This photo shows several different ways that fiber optic cables are bundled together. The top cable contains 20 individual fibers. The second cable combines fiber and traditional copper cables. The bottom cable is the most common type, a single fiber cable.

Cable Construction This is how a single fiber cable is constructed. Keep in mind that the fiber itself is actually a very long piece of high-quality glass. It is coated, then placed into a buffer. The Kevlar yarn is extremely strong, and provides that cable with the toughness to survive normal handling.

Fiber end-view Here’s a single fiber cable viewed from its end. It is constructed very much like a coaxial cable, but with different materials. The glass fiber is the center of the cable. The glass is surrounded with a material called cladding, that adds strength. Surrounding that is the buffer. The buffer is similar to the insulation on a copper wire, although it performs a different function. Then finally, the jacket.

The Glass Fiber This is an extreme close-up of a fiber cable. To give you some idea of the size, that’s the tip of a thumb. You can clearly see the white buffer. Further down the cable you can see the cladding, then at the end is the actual glass fiber. This fiber is smaller than the average human hair.

ST connector You won’t likely be handling the glass fibers themselves, they are normally housed in a connector such as this ST connector. There are many different types of connectors for fiber cable, this is one of the more common types. If you are familiar with a BNC connector, it works the same way.

ST Connector Finally, this is where the ST connector is attached. It resembles the BNC connector but is about half the size.

LAN Communication

LAN Technologies: Ethernet Token Ring ARCnet LocalTalk
Over the years several different networking technologies have been developed. These include Ethernet, Token Ring, ARCnet, which is being phased out in favor of faster technologies; and LocalTalk, which is integrated right into Apple’s Macintosh computer. This is by no means a complete list and you will hear of other technologies.

Ethernet Developed by Xerox in early ’70s.
Has become most popular networking technology in use today. A variety of speeds and cabling options have evolved. It is fast, inexpensive, and flexible. It continues to evolve. Majority of new networks use Ethernet. Ethernet is by far the most popular LAN technology in use today. Originally developed by Xerox in the early 1970s, it has gone through several evolutions. Today, several different speed and cabling options are available, depending on your needs and your budget. Ethernet has always been relatively fast, inexpensive and flexible compared to the alternatives. It has continued to evolve over the years with faster versions becoming available as needed. Its low cost, ease of installation, and flexibility make it the technology of choice for most new networks.

Networking is all about sending data from one location to another.
May I have Worksheet-101? Sure, here it is! As we have seen, networking is concerned with sending data from one location to another. Here one computer requests a worksheet. The other computer provides it. Easy, right? PC-2 PC-1

The CPU routinely sends data from one place to another.
int e l After all, data is sent from one place to another inside the computer all the time. The CPU routinely communicates with memory, the keyboard, the floppy disk, the hard drive, etc. Can sending data to another computer be all that difficult?

But when multiple computers are involved, a new series of problems arise.
PC-1 PC-2 PC-3 PC-4 PC-5 Well, it turns out that when multiple computers must communicate via a single path as they do in a typical network, a whole new family of problems arises.

How do you keep all the computers from transmitting at the same time?
For example, how do you keep all the computers from transmitting at once? On most networks, if even two of the computers transmit simultaneously, both transmissions are garbled and no data gets to its intended recipient. Some system has to be developed that allows orderly access to the cable. PC-1 PC-2 PC-3 PC-4 PC-5

Is this for me? PC-1 PC-2 PC-3 PC-4 PC-5
Once you have solved that problem and the computers are transmitting in an orderly manner, how do the other computers know who the transmission is for?

Who sent this? PC-1 PC-2 PC-3 PC-4 PC-5
And once you do get the data to the proper machine, how does that machine know who sent the data? After all, it may want to report that it received the data properly. But, who does it send the message to?

Gee, I wonder if it is correct.
Ah, it’s from PC-1. Gee, I wonder if it is correct. PC-1 PC-2 PC-3 PC-4 PC-5 Just as important as knowing who the transmission is from is knowing whether or not the transmission was received properly. If an error occurred during the transfer, you certainly want to know about it. So how do you handle error detection? And if an error is detected, what do you do about it?

Here’s the ten gigabyte folder you wanted!
PC-1 PC-2 PC-3 PC-4 PC-5 Another problem that is not so obvious is this one. How do you keep a single computer from dominating the network, shutting out the other computers for long periods of time? This situation is compounded by the fact that the longer the transmission, the more likely it is to contain an error. This means that longer transmissions are more likely to have to be retransmitted, tying up the network for even longer periods. Let’s talk about this problem first because its solution is fundamental to virtually all networks.

Large files are broken into manageable chunks called packets.
The solution to this problem is both simple and elegant. Instead of sending data as a single large file, the file is broken up into small manageable chunks called packets. A one megabyte file might be broken up into hundreds of small packets. This has two great advantages. First, it gives the other machines on the network a fair chance of grabbing access to the cable in between packets. Second, if errors do occur during the transmission, only those packets that contain an error need to be retransmitted. This simple technique has proven to be so effective that it is used on virtually all networks. Packets

How do you keep two computers from transmitting at the same time?
Now let’s consider the other problems. Of all the problems mentioned perhaps the thorniest is: “How do you keep two or more computers from transmitting at the same time?” PC-1 PC-2 PC-3 PC-4 PC-5

Carrier Sense Multiple Access/ Collision Detection (CSMA/CD)
Ethernet’s solution to this dilemma is CSMA/CD. In fact, Ethernet is virtually synonymous with Carrier Sense Multiple Access with Collision Detection. Let’s see what all this means.

Carrier Sense Each computer attached to the network examines the cable before transmitting. If it senses traffic on the cable, it waits until the traffic clears before transmitting. First, carrier sense simply means that each computer on the network can detect traffic on the cable by “listening” to the cable. Any computer that wishes to transmit over the network listens first to make sure the cable is not in use. If the computer finds the cable idle, it goes ahead and transmits its data. On the other hand, if it finds the cable in use, it backs off until it finds the cable free again. This can be compared to listening for a dial tone before dialing the telephone or looking both ways before pulling out into traffic. It is a simple technique, but it goes a long way toward solving the media access problem.

Multiple Access All computers on the network have equal access to the cable. A lowly desktop has the same access as the Windows NT Server. Access is on a first-come, first-served basis. The only consideration is: “Is someone else using the cable?” Multiple Access simply means that all computers on the network have an equal shot at using the cable. Ethernet is completely democratic in this regard. The lowliest desktop has the same access as the highest server. Access is strictly on a first-come, first-served basis. The only consideration is: “Is the cable free?” If so, the computer transmits its packet.

Collisions still happen
In spite of these precautions, quite often two computers will listen, hear nothing, and then transmit at the same time. When two computers transmit at the same time, both transmissions are garbled, and neither gets through. This is called a collision. This is where the collision detection or “CD” part of CSMA/CD comes in. PC-1 PC-2 PC-3 PC-4 PC-5

The computers involved in the collision detect the collision.
Wow! That’s not what I said! Wow! That’s not what I said! When a collision occurs, the two computers involved detect the collision. They do this by comparing the packet they sent with the packet on the cable. When the machine sees a difference between what it sent and what ended up on the cable, it assumes that a collision has occurred. PC-1 PC-2 PC-3 PC-4 PC-5

The computers involved in the collision back off for a random length of time.
I’ll try “y” Microseconds. I’ll back off for “x”Microseconds To prevent a repeat of the collision, the two computers involved back off for a very short but random length of time. The idea is to make the delays different so that the packets will not collide when retransmitted. Here let’s assume that PC-1 has the shorter delay. When its delay “times out”, PC-1 will listen for activity on the cable. If the cable is free, it will go ahead and retransmit its packet. More than likely, when PC-2’s delay times out an instant later, the cable will be in use and PC-2 will have to wait for the next opportunity to transmit. In this way, the computers are forced to take turns, but eventually both will transmit their packets. PC-1 PC-2 PC-3 PC-4 PC-5

Collisions are a natural characteristic of Ethernet.
Collisions are a natural characteristic of Ethernet. Retransmitting the colliding packets slows the overall network. However, this simple method of resolving media contention results in circuitry that is simple and inexpensive. The low cost of the Ethernet has been one of the main factors driving its popularity. PC-1 PC-2 PC-3 PC-4 PC-5

Unit 10 Networks.

Similar presentations

Presentation on theme: "Unit 10 Networks."— Presentation transcript:

Similar presentations

About project

Feedback

Log in

Auth with social network:

Unit 10 Networks.

Similar presentations

Presentation on theme: "Unit 10 Networks."— Presentation transcript:

Similar presentations

About project

Feedback