1. Create several duplicate scenarios and modify the interarrival times for all the Ethernet stations to 0.0008, 0.002, 0.003, 0.005, and 0.006, respectively. Rerun the simulation, and record the Traffic Received (bits/sec) statistic for each scenario. Using a spreadsheet, plot the values, (together with the results you found in the preceding scenarios using interarrival times of 0.001 and 0.004). Find the interarrival time for which the maximum throughput is achieved. Calculate the offered load that corresponds to this interarrival time. Why does the throughput stop increasing even when the load is increased further?

2. Rerun the simulation using the same interarrival times in Question 1. For each run, record the utilization and Traffic Received (bits/sec). Plot the values using a spreadsheet. How are the two statistics related? Why can't the system reach 100% utilization?

3. Some interactive applications require very short delay. Excessive load and resulting collisions can increase delay significantly. Modify the interarrival time and rerun the simulation to determine the minimum interarrival time and corresponding per-node load if the delay must be 0.005 second or less.

4. Modify the interarrival time and rerun the simulation to determine how large the interarrival time must be in order for the number of collisions per second to become negligible (less than 10 per second). What is the pernode load that corresponds to the interarrival time you found?

5. Under high loads, shared Ethernet LANs tend to perform poorly. Examine the Delay statistic for the High_Load scenario again. How does the delay experienced by packets change over the run of the simulation? Why does this behavior occur? What will happen to the delay if the load continues at the same level?