Networking in a Nutshell

Networking is a means of connecting computers together so that they can share data, such as files and programs, and also resources, such as printers, servers, and Internet connections. As with all other computer related subjects, networking can be separated into two categories: network hardware and network software.

Network design includes the physical components, network technologies, interfacing software and their protocols needed to establish reliable communication among each computer or workstation, which is called a node, or a host. Data is sent over a network as bits and bytes that have been translated into electronic signals or light bursts in the case of fiberoptic networks. Before being transmitted, data is first divided into segments, each of which has a header and trailer attached. The headers and trailers are called a frame and the entire unit including the data is called a data packet or packet.Each data packet is sent as an independent unit over the network. At the receiving end, the header and trailer information is removed, and the data within the several packets is reassembled in order into the original complete data. Obviously this requires a tremendous amount of overhead or header information. Sometimes the size of the header/trailer is larger than the data in a given packet.

You can think of this process of sending data over a network as similar to that of shipping your computer to your friend in California. The computer is first disassembled and packed into several boxes. When the boxes all arrive, the components are unpacked and reassembled into a complete computer. Think of the computer as data, its various components as data segments, and the shipping boxes with address labels as the headers and trailers. In addition, there are some large Post Office containers and trucks that are part of the shipping process. These would be considered headers also as they are part of the overhead that gets the boxes delivered. And finally, there are some things you can see(software?) such as Postal Service rules and protocols that determine how the boxes get shipped. These protocols would be considered headers also. The units (the combination of shipping boxes, labels, and computer components, Post office containers and protocols) are the data packets. While in transit, the computer (or data) cannot function as a computer because it has been temporarily disassembled. In fact, the individual boxes(packets) may not even be shipped on the same postal truck or arrive on the same date. But once it arrives at its destination, the packing is removed, and the computer is reassembled and a done deal.

A header is information sent in front of data to identify the data destination and the protocols that the packet is using. A trailer follows the data and contains information used by some protocols for error checking. In PCs network interface card(NIC) breaks the data into segments and encloses each segment between headers and trailers, thus creating individual packets. In addition, on the receiving end, the firmware on the network card reassembles the segments back into contiguous data. In the computer shipping analogy, think of the network card in the sending PC as you as you dissassembled, packed and shipped the components. Think of the network card in the receiving PC as your friend who receives the boxes, unpacks them, and assembles the computer system.

And finally, there are two approaches to managing a network: a peer-to-peer network where each individual workstation manages its own security and resources, and a client-server network that is managed by a server. The majority of large networks are client-server. At school, you log into the network server to “get on the network”. The server recognizes your login name of “student” and gives you access to certain resources within the network. If you login on the same machine using my login name and password, you get different network services. This client-server approach allows easy management of large groups of users. In home networking, you might prefer to set up a peer-to-peer network to save the cost of the server since you only have a few nodes. The Cisco Curriculum talks almost completely about large managed client-server networks. You will have some chances to design such a network in this class. You will also be given an opportunity to design a small peer-to-peer network in this class.

The OSI Layer Network Model

In the 1970s, when manufacturers were beginning to build networking software and hardware to connect computers, each manufacturer developed its own standards of communication within its proprietary network design. In the early 1980s, manufacturers began to make attempts to standardize networking so that networks from different manufacturers could communicate. Two bodies that were leaders in this standardization are the International Organization for Standardization (ISO) and the Institute of Electrical and Electronics Engineers (IEEE). For example, one major effort of the IEEE was to standardize Token Ring and Ethernet protocols, which are both considered industry standards for network cards and other network hardware devices that together make up the physical devices that form a network. Ethernet and Token Ring networks are discussed later in the Cisco Curriculum.

In an overall effort to identify and standardize all the levels of communication needed in networking, ISO developed a networking model called the Open Systems Interconnect (OSI) reference model, which is illustrated at the bottom of this section. This model breaks down the communication needed for one user to communicate with another over a network into seven logical levels. These individual layers each add overhead to the data that becomes the headers that you saw in the first section of this reading. Basically as you send a packet to another node, the encapsulation(adding of headers) starts at Layer 7 in your computer and goes 7-6-5-4-3-2-1. Each layer adds some overhead information in headers until it leaves your computer at Layer 1 and travels along the wire through the network. When it reaches the destination, it goes from the wire: 1-2-3-4-5-6-7 with each layer of the receiving computer reading the layer header and removing it. The header in a given layer tells the receiving computer where and how to pass the packet up to the next layer within the computer until it reaches Layer 7 and bursts on to the recipients computer screen as a webpage or an email or whatever. Communication between adjacent layers is considered direct, but communication between matching layers is considered logical or virtual.

When studying the OSI model, remember that not all networks or packets have a separate software layer that matches each of the seven layers. Realistically, no network in use today perfectly follows the model. However, the model does serve the networking industry as a reference point for discussing different levels or layers in a network. For example, the network card operates in the physical layer(Layer 1) and the data-link layer(Layer 2) in the model. From the image below we see that the data-link layer is responsible for disassembling data(including upper layer headers) into segments to be assigned to separate packets and later reassembling packets into contiguous data. The physical layer is responsible for passing packets to and receiving packets from the network media or cabling.

The third layer that we will really focus on in this class is the Network Layer(Layer 3). Once you know that the NIC on an Ethernet network(or Token Ring) handles the bottom two layers of data transmission, you can ask the question, "What software on the network determines the best possible route to send a packet so it can arrive at its destination?" This question is addressed by the network layer. The network card(NIC) of the Ethernet network has nothing to do with “best path determination”, because Ethernet does not encompass the network layer of the OSI model. By referring to the OSI model, students (and network engineers) have a structure from which to work as they understand what is going on in a given network.

Without getting too deeply into the details of the OSI model, the following provides an overview of the role each layer plays in a network, starting at the top.

Application Layer(7)

The application layer of the OSI model is responsible for interfacing with the applications software using the network. For example, suppose you are using a word processor such as Microsoft Word. Word can open a document file that is stored on your hard drive (drive C), and it can just as easily open a document file stored on a file server connected to the LAN. The file server is known to the word processor as just another drive, such as drive F. You can open C:\Myfile.doc or F:\Myotherfile.doc. When the word processor attempts to open the file on drive F (which is the file server), it virtually communicates the request to the application layer of the network software. We say virtually because your computer first has to follow the OSI model down(7-6-5-4-3-2-1) adding headers along the way. The entire request data packet(notice in this case the data is just a request for services), travels as bits along the network to the server. The server reads and removes the headers to understand your request. It then packages the file you requested(F:\Myotherfile.doc) with appropriate headers to get it along the network back to you. The file is transmitted over the network and presented to the word processor by the OSI application layer(Layer 7).

Presentation Layer(6)

The presentation layer is in charge of reformatting, compressing, or encryption of data in order for the application layer above it and the session layer below it to communicate(reformatting), for the data to be sent faster(compression), or to secure the data(encryption).

Session Layer(5)

The session layer is responsible for establishing and maintaining a session between two networked stations or nodes. A session over a network works somewhat like a telephone call over phone lines. The caller makes a call; someone answers on the other end. After both parties know that communication is established, conversation goes in both directions until either the caller or receiver ends the phone call. The session layer performs similar duties. An attempt is made to establish a session between two nodes on a network. Both nodes acknowledge the session, and the session is usually assigned an identifying number. Either node can disconnect a session when communication in both directions is completed. Sometimes a session between two nodes on a network is called a socket. When a session is established, a socket is opened. A disconnected session is called a closed socket. Note that the session layer is not in charge of sending data, determining the path data takes, or creating the data packets. So in our phone call analogy, just the dialing/ringing of the phone and the “hello”, “hello yes it’s me let’s talk” would be analogous to the session layer.

Transport Layer(4)

The transport layer is responsible for error checking and requests retransmission of data if errors are detected. The transport layer guarantees successful delivery of data. We should note that some data packets do not add an error checking/retransmission header as it cuts down on the amount of overhead(file size) in the packet

Network Layer(3)

The network layer is responsible for finding the best possible route by which to send frames over an internetwork (a network of networks). The two most common protocols that make up both the transport layer and the network layer are TCP/IP (Transmission Control Protocol/Internet Protocol) and IPX/SPX (Internetwork Packet Exchange/Sequenced Packet Exchange), which are both supported by Windows NT, Windows 2000, and Windows 9x. For TCP/IP, which is used by the Internet, the TCP portion of the protocol is responsible for error checking, and therefore operates in the transport layer. The IP portion of the protocol makes up the network layer and is responsible for routing. IPX/SPX is used by NetWare by Novell, one of the most popular network operating systems for LANs. Novell also now supports TCP/IP and the IPX/SPX protocols may be going the way of the dinosaur. The IPX portion of the protocol is the network layer responsible for routing, and the SPX portion of the protocol manages error checking, making it the transport layer. These network protocols are further discussed later in the Cisco Curriculum.

Data-Link Layer(2)

The data-link layer is responsible for receiving data from the network layer and splitting them up into segments of bits(data packets) to be presented to the physical layer for transport. When the bit stream is received from the physical layer, the data packets are reconstructed to be presented to the network layer. Token Ring and Ethernet network cards are examples of devices that handles both the data-link and physical layers of the OSI model.

Physical Layer(1)

The OSI physical layer on a PC is controlled by the network card. It includes the IEEE specifications for cabling types connected to the card and other media definitions. This layer controls how data is transmitted over the physical media. At this level, data is nothing but indistinguishable bits. Remember that data is packaged within frames before it is transmitted. This packaging of data has already occurred before this layer(gradually in the upper layers and finally in Layer 2); the physical layer does not distinguish the frame header or trailer from the payload, or data, within the frame. The physical layer sees all of it as just bits that need to be passed on.

One Last Summary of the Networking Model

Remember that data is segmented and enclosed in frames before it is transmitted over a network as data packets. Each layer in the seven-layer model can add information to the beginning and ending of a data packet to be read by the counterpart layer on the receiving workstation. In practice, however, only the data link layer adds both a header and a trailer. The physical layer adds neither, and the other layers might or might not add a header. At its most complex stage, a packet may look like the image below, in which each layer that adds a header or trailer has added its identifying information to the packet.

Later, when the packet is presented to the counterpart layer on the receiving station, that layer interprets any information in the header and trailer intended for that layer. Then it strips off that header and trailer and passes the packet to the next higher layer in the model. For packets to transmit successfully, each layer of the OSI model must communicate using the same protocol as its counterpart layer on the remote computer.


Index Class Schedule Cisco Curriculum students.netacad.net Class Assignments Class PowerPoints