A Web-Based Introduction to Computer Networks for Non-Majors

Media Access in Ethernet

A Laboratory

The Ethernet applet and the accompanying notes (such as How do I use the applet? and What concepts is this applet illustrating? ) can be used as the basis of a series of laboratory experiments as shown below. Each experiment has a series of questions which should be answered as part of the experiment.

The type of network used to connect computers that are close together (such as in the same building) is said to be a local area network (LAN). Ethernet is a popular form of local area network. For example, the local area networks in the buildings on many campuses are Ethernets. An Ethernet can be thought of logically as having the topology of a bus. Each machine (also called a station) is connected to the bus, but only one machine can be sending a message (also called a frame) on the bus at a time. The Ethernet bus is also called the media because the bus is the means by which the machines communicate (just like newspapers, television, and radio are called media since they are forms of communication). Since only one machine can use the bus at a time there must be some set of rules to determine which machine gets to use, or access, the bus if more than one wants to. This set of rules, or protocol, is called the media access protocol of Ethernet. See the document What is Ethernet's Media Access Control Protocol (CSMA/CD)? for a description of this protocol. This laboratory complements that document be being an alternative method of investigating the protocol. We will call the Ethernet bus, a link, to be consistent with the common Ethernet terminology.

As described in the document What is Ethernet's Media Access Control Protocol (CSMA/CD)?, the time for a frame to reach a destination is divided into the propagation delay and the transmit time. This experiment explores the difference between these two time periods and how each is influenced by the basic properties of an Ethernet (link length, link bandwidth, and frame size).

Initially, place one station on the link which is the Ethernet bus and enter the minimum frame size, minimum link bandwidth, and minimum link length into the text boxes. Click on the station and watch its frame copies propagate to the ends of the link. Increase the frame size and repeat the experiment. Keeping the link bandwidth and link length constant, change the frame size and repeat the simulation. Keeping the link bandwidth and frame size constant, change the link length and repeat the simulation. Keeping the link length and frame size constant, change the link bandwidth and repeat the simulation.

1. Is a frame’s propagation delay a function of the link length? of the link bandwidth? of the frame size?
2. Is a frame’s transmit time a function of the link length? of the link bandwidth? of the frame size?
3. If link length is long and frame size is small, which of propagation delay and transmit time form the largest fraction of the total time for a frame to reach its destination?
4. If link length is small and frame size is large, which of propagation delay and transmit time form the largest fraction of the total time for a frame to reach its destination?
5. Graphically how does the appearance of the rectangle that represents a copy of a frame change as the link length increases? as the link bandwidth increases? as the frame size increases?

Warning: To make the simulation as visible as possible, changing the link length just changes the numbers on the tick marks on the bar representing the link. The actual size of the bar on the screen is not changed. A frame copy always moves along the link at about two-thirds of the speed of light. Consequently, if the link length increases, the simulation should have each frame copy appear to move more slowly towards the link end (since the speed of the frame copy has not changed, but the link length has increased). The simulation does not try to change the frame’s rate of moving across the screen as the link length changes. Thus, in deciding if the propagation delay is a function of link length you can not just watch the rate the frame is moving across the screen.

In experiment one since there was only one station, that station went through the sequence of protocol states that indicates a successful transmission. When there are two or more stations on the link, complications may arise. This experiment investigates one complication. In this complication when a station is starting to attempt a transmission, a frame copy from a second station is passing in front of the first station.

1. Place two stations on the link. Click on just one station to start it transmitting a frame. What is the sequence of protocol states that the station goes through before the frame transmission completes?
2. Place two stations on the link. Click on one station to start it transmitting a frame. Now, while a copy of the first station’s frame is passing by the second station, click on the second station. What is the sequence of protocol states that the second station goes through? Explain why this sequence differs from that sequence the first station followed.
3. Place two stations on the link. Click on one station to start it transmitting a frame. Now, do the click on the second station before the first station’s frame copy reaches the second station. What is the sequence of protocol states that the second station goes through? Explain why this sequence differs from that sequence the second station followed in the immediately preceding question.

Experiment two considers one complication that can arise when two or more stations are on the link and attempting to transmit. This experiment examines another complication. This complication involves the protocol behavior when a station detects that a frame it is transmitting has or will collide with a frame from another transmitting station. Place two stations on the link. Click on one station to start it transmitting a frame. Click on the second station to start it transmitting a frame before the frame of the first station reaches the second station; thus, the complication described in experiment two will not occur. Initially set the simulation parameters (frame size, link bandwidth, and link length), the position of the stations, and the timing of the clicks so that each station has finished emitting its frame before the frame copy of the other station reaches it.

1. Describe what happens to the frame copies of the two stations. Explain when each frame copy changes state.
2. Describe what happens to each of the stations. Explain when each station changes state.
3. Now modify the timing of the clicks so that each station has not finished emitting its frame before the frame copy of the other station reaches it.
4. Describe what happens to each of the stations in this case. Explain when each station changes state.
5. What is the meaning of the number that appears below each station’s rectangle? The number appears below a station’s rectangle only after that station has detected that its frame is involved in a collision.

While a station is transmitting a frame, that station will notice if any other station’s frame passes in front of it. If this happens, the station realizes that its frame has collided with that other frame. This is the only way a station detects a collision. It is essential that a station sending a frame detect whether that frame is involved in a collision. Thus, a station must continue transmitting a frame long enough that the station is guaranteed to always detect a collision. The minimum transmission time that is "long enough" is called the slot time. This experiment investigates how long is "long enough".

Initially place two stations near each other on the link and use the minimum frame size, minimum link bandwidth, and minimum link length. Click on the two stations in rapid succession and observe the frames collide. Both stations should detect the collision. If either of the stations does not detect the collision, then place the stations closer together or make the time between the click on each station smaller. Since both stations are detecting the collision, then the minimum frame size causes a long enough transmission time for this situation. However, what if the placement of the stations or the timing of the clicks changed? The slot time is the transmission time that is long enough regardless of where the two stations are placed or when the clicks occur.

1. Change the placement of the two stations. Is there a placement of the two stations that causes the minimum frame size to be too small? In other words, is there a station placement that causes the minimum frame size to cause the transmission time to be so small that a station might not detect a collision? If so, what are the positions of the stations (you should be using the minimum link bandwidth and minimum link length)?
2. If there is a placement of the stations that causes the minimum frame size to be too small, then increase the frame size and repeat the experiment. Continue repeating the experiment until you find a frame size that is large enough that both stations will detect a collision. Is there a placement of the two stations that causes the new frame size to be too small? If so, what are the positions of the stations (you should be using the minimum link bandwidth and minimum link length)?
3. As the placement of the two stations changes, the smallest frame size large enough so that both stations are guaranteed to detect a collision, will change. How does it change? What is the general rule for how station placement effects the smallest frame that is large enough?
4. Which station placement has the largest smallest frame size that is large enough?
5. So far you have just been changing the placement of the stations, but you have continued to click on both stations in rapid succession. Now increase the time between the click on the first station and the click on the second station. How does increasing this time change the size of the smallest frame that is large enough so that both stations detect a collision?
6. What station placement and rule for when to click on each station causes the largest smallest frame size that is large enough so that both stations detect a collision? The time it takes a station to emit this largest smallest frame is the slot time.
7. Does adding more stations change the slot time? Why or why not?

In Experiment Four for a fixed link bandwidth and link length you found the slot time; that is, the largest smallest frame that is large enough so that both stations are guaranteed to detect a collision regardless of station placement and when each station is clicked on. This experiment investigates how changing either the link bandwidth or the link length effects the slot time.

1. For a given link bandwidth, as the link length increases, what happens to the slot time? Explain.
2. For a given link length, as the link bandwidth increases, what happens to the slot time? Explain.
3. The original version of Ethernet used in the buildings on our campus had a link bandwidth of 10 Megabits per second and a maximum link length of around 1500 meters. The current version of Ethernet, called Fast Ethernet,has a link bandwidth of 100 Megabits per second. What do you think is the maximum link length of Fast Ethernet? Explain.