Class Assignments

The two bottom levels of the OSI model, the data-link layer and the physical layer are more directly connected to the physical hardware devices that form a Local Area Network(LAN) than other layers in the OSI model. The type of Network Interface Card that operates at the data-link layer and physical layer is determined by the physical technologies used to create the network. These physical technologies and related network hardware devices are sometimes call LAN Technologies. Near the end of this class we will cover Wide Area Network(WAN) Technologies. WAN technologies cover how the data leaves your building or home and enters outside of your local network. Network Administrators are concerned with WAN connections and how data travels from one building(LAN) to another but this section will look at the technologies that allow a single LAN(usually a building or a section of a very large building) to communicate with itself.

The three most popular LAN network technologies are Ethernet, Token Ring, and Fiber Distributed Data Interface (FDDI). A relatively new type of network is Asynchronous Transfer Mode (ATM), which is a very fast network that works well over both short and long distances. Each type of technology is designed to solve certain network problems, and each has its own advantages and disadvantages. Network technologies differ from each other in many ways. The coverage of some detailed differences is beyond the scope of this class. However, we will briefly discuss the three main LAN technologies.

Ethernet

Ethernet is the most popular LAN technology used today. Ethernet networks can be configured as either a bus topology or a star topology. Topology is the arrangement or shape used to physically connect devices on a network to one another. Figure 17-3 shows an example of a bus and a star topology. A bus topology connects each node in a line and does not include a centralized point of connection. Cables just go from one computer, to the next one, and the next. A star topology connects all nodes to a centralized hub. PCs on the LAN are like the points of a star around a central hub. Of these two topologies, the star(and its hybrid, the extended star) are the preferred topologies in most ethernet LANS.

The star arrangement is more popular because it is easier to wire and to maintain than is the bus arrangement. In a bus arrangement, the failure of one node affects all the other nodes. In a star arrangement, the only node that can shut the network down is the center of a star, usually a switch or a hub.

An Ethernet network is a passive network, meaning that the networked computers have equal access to the network. There is no “controlling device” that determines who sends data when. You might even call it “networking anarchy”, except that there are rules and every node is polite and follows them. Ethernet works much like an old telephone party line, where each computer is like a party line caller. When someone on a party line wanted to use a phone, he or she would pick up and listen. If there was a dial tone (carrier), then the person could make a call. If someone else was talking, the person would hang up and try again later. If two people attempted to make a call at the same time, both calls would fail. They would each need to hang up and begin again. The first one back on the line would be able to make a call.

Similarly, a computer that wants to send packets over Ethernet will first listen on the network for silence(above diagram). If it hears nothing, it begins to transmit. As it transmits, it is also listening. If it hears something other than its own data being transmitted, it stops transmitting and sends out a signal that there has been a collision, which occurs when two computers attempt to send data at the same time. A collision can corrupt packets the computers recently sent. Each computer waits for a random amount of time and then tries to transmit again, first listening for silence. Computers using Ethernet are said to gain access to the network using the Carrier Sense Multiple Access with Collision Detection (CSMA/CD) method. The name of the method suggests three characteristics of the way computers communicate on Ethernet: (1) a computer must sense that the network is free to handle its transmission before initiating a signal (carrier sense), (2) many computers use the same network (multiple access), and (3) each computer must detect and manage collisions (collision detection). This type of LAN technology is ideal for networks where you do not have regular, same-every-day traffic. A network in which anyone could use it at any time, but might not. In other words, a typical office or computer labe environment. See, a polite anarchy… with rules. Kind of… Sort of.

Ethernet can use any one of six cabling systems, including 10Mbps and 100Mbps versions of twisted pair cable, fiber optic, or coaxial. The two most popular Ethernet cabling systems are 10BaseT and 10Base2 (Thinnet). The image below shows an example these two cables.

The "10" in 10BaseT comes from the speed of transmission (10 Mbps). The "Base" comes from baseband. Ethernet is a baseband network, which carries data over wire a single message at a time in digital form. Contrast baseband to a broadband network such as ATM or cable modem that carries multiple messages over wire, each message traveling on its own frequency in analog form. Note: The term “broadband” is misused today by many companies seeking customers. The "T" in 10BaseT stands for twisted-pair cabling. 10BaseT networks use RJ-45 connectors, which look like large phone jack connectors.

The slightly less popular Thinnet (10Base2) networks use coaxial cables and BNC connectors, which are sometimes shaped like T's. Thinnet networks use a bus topology with terminators at each end of the bus that twist into the T-connectors at the back of the two end PCs. The BNC port on the network card looks like a cable TV connection.

Because signals transmitted over long distances on a network can weaken, devices are added to strengthen the signals. For example, for a 10BaseT Ethernet cable, if the cable exceeds 100 meters(328 feet), signal strengthening is required. A repeater is a device that strengthens signals on a network. The repeater recreates the original signal by amplifying it and retiming it. Without a repeater, the signal becomes weak and loses its defining wave pattern.

Each of the cable systems listed in the table can support only a limited number of nodes. As the number of nodes increases, performance speed and reliability can drop for the overall network. One method used to prevent this kind of congestion is segmentation. Segmentation splits a large Ethernet into smaller Ethernet segments. Each segment contains two or more computers and is connected to the other segments by a router, switch, or bridge(Note: A hub is a “dumb” device and will not segment). The differences among these devices are discussed later in this course. For now, know that routers, switches, and bridges allow Ethernet LANs to be broken up(segmented) into smaller groups of workstations. Think of it like our school. At any given hour we have segmented students into individual classrooms. This allows more effective communication. The math classroom can talk math, the english classroom can talk english, and the Intro to Networking class can talk Ethernet. If we were all in one large room, say the gym, teaching/learning would be very difficult. However, we are all connected. Using the proper protocols, say a hall pass, you can leave the Intro to Networking room and travel down the hall to a math room. However, you only do that when you need to connect with a math teacher. Most of the communication in the Intro to Networking classroom stays within our classroom and does not interfere with other building conversations. This is the function of Ethernet segmentation, to keep local traffic within each segment and forward non-local traffic to the correct segment.

Token Ring

Token Ring networking, which was developed by IBM, is more complex and expensive, but more robust and reliable, than Ethernet. Because of its complexity, it is more difficult to maintain than Ethernet. Its reliability also makes it perfect for networks that have very standard traffic patterns. One use of Token Ring is in manufacturing. Its reliability allows you to time tasks within the factory so that manufacturing automation(robotics) is possible.

Logically, Token Ring networks are rings of stars. Each star is a segment on the ring. Each workstation connects to a centralized control device called a multistation access unit (MSAU). Already you can one big difference between Ethernet and Token Ring. Ethernet was a free for all, Token Ring has controllers(MSAU’s). One MSAU can connect to another by a cable called a patch cable. One end of each MSAU has a Ring In (data flows into the MSAU) or Ring Out (data flows out from the MSAU connection. The main ring is composed of the MSAUs and the cables connecting them, which are together referred to as the main ring cable. The main ring cable today is often fiber-optic.

The entire token ring is made up of not only the main ring, but also the cabling to each PC on the token ring. Each workstation contains a Token Ring NIC card with a 9-pin connector for the Token Ring cable, which connects each workstation to an MSAU. Each Token Ring network card has a unique address like an Ethernet card, which is assigned to it during manufacturing. Token Ring cables can be either UTP or STP cables that have two twisted pairs, or four total wires in the cable. Looking at the diagram below, you can see why it is a ring topology despite the stars off of the MSAU’s. Traffic flows in one direction around the main ring, usually at a defined, constant rate.

Communication and traffic on a Token Ring network are controlled by a token, which is a small frame(Layer 2 “data packet”) with a special format that travels around the ring in only one direction. One station receives the token from the preceding and passes it on to the next station on the ring. As one station passes the token to the next station, it can attach data in a frame to the token. The next station receives the token together with the data frame and reads this data frame. If the frame is intended for it(by reading the hardcoded destination MAC address), it changes 2 bits in the frame to indicate that the data has been read by the intended station. It then passes the token and the data frame on. When the token and frame are received by the station that sent the frame, it sees that the frame was successfully received and releases the token by passing it on to the next PC, without a data frame attached. However, if the amount of data requires more than one frame, instead of releasing the token, the PC sends the segment of data with the token. In either case, the token is passed on to the next PC, and data is never on the ring without the token attached to it.

Any PC receiving a token with no data frame attached is free to attach a data frame before passing on the token. The token is busy and not released to another PC until the sending PC has received word that the data was successfully received at its destination. In other words, the only PC that should remove a data frame from behind the token is the PC that attached it in the first place. One of the real advantages of token ring networks is that there can be no collisions since there is only one token and only one node can send at a time. You might think this is “slow” way of networking but remember electronic traffic moves fast and in theory ethernet also should have only one sender at a time.

FDDI

Fiber Distributed Data Interface (FDDI, pronounced "fiddy") is a ring-based network, like Token Ring, but does not require a centralized hub, making it both a logical and physical ring. FDDI provides data transfer at 100 Mbps, which is much faster than Token Ring or regular Ethernet, and a little faster than Fast Ethernet, which also runs at about 100 Mbps. At one time, FDDI used only fiber-optic cabling, but now it can also run on UTP. FDDI is often used as a backbone network. A backbone is a network used to link several networks together. For example, several Token Ring and Ethernet networks can be connected using a single FDDI backbone.

FDDI uses a token-passing method to control traffic, but FDDI is more powerful and sophisticated than Token Ring. FDDI stations can pass more than one frame of data along the ring without waiting for the first frame to return. Once the frames are transmitted, the sending station can pass the FDDI token to the next station, so more than one station can have frames on the ring at the same time. With Token Ring, a data frame is only found traveling behind the token. With FDDI, data frames travel on the ring without the token. A PC keeps the token until it has sent out its data and then passes the token on. Possessing the token gives a PC the right to send data. A token is released (sent on) when the PC has finished transmitting

One important strength of FDDI is its dual counter-rotating rings, as seen in the diagram. Instead of a single ring like the one the Token Ring uses, FDDI has two rings linking each device on the network, a primary ring and secondary ring. Data normally travels on the primary ring. However, if a break occurs on the FDDI ring, any device can switch the data to the secondary ring, which causes the data to travel in the opposite direction back around the ring, as shown in the above diagram. When the data reaches the break coming from the other direction, a station switches the data back to the primary ring, and it continues in the opposite direction again. In this way, communication continues even with a break in one FDDI ring.

Index