Exercise 8: Peak-driven traffic model for data

Now we will look at the Internet Access service and compare how voice and data services are dimensioned.

1. Return to the Demand dialog for all three services and look at the column for Internet Access. The penetration for this service is constant at 1.0, meaning that all subscribers will use this service.
2. Notice that Traffic Calculation is set to Peak Driven. This means that the calculation starts with the busy-hour traffic assumptions and works up to the annual traffic (volume).
3. Look at the inputs for Nominal Bandwidth per Connection and Contention Ratio (over-subscription factor). These assume a nominal 2Mbit/s service, but peaking at an average of 1/50 of that over all subscribers.

What factors might influence the contention ratio? Think about how you use applications like email and Web browsing, how intensively you use these applications, and what proportion of your time might be away from the computer.

4. Review the formula for Annual to Busy-Hour Unit Ratio and try to explain each of the factors involved.
5. See if your explanation is consistent with a Traffic Unit of GBytes!

The annual traffic is calculated from the assumptions below and here the Traffic per Connection input for Internet Access is disregarded.

6. Draw the following graphs and see if they match your expectations. (Try to work out the orders of magnitude in your head first.)

• Busy Hour Traffic (Internet Access)
• Connections (Internet Access)
• Traffic (Internet Access)

Try changing the assumption for Contention Ratio to see how it affects the results.