PHASE 3 : Develop and Test the Server class

A complex project like this needs to be developed in phases. In the third phase, you will define a class to represent a server for the simulation framework and test it using a provided test driver.

Implementation Notes

1. Here is a ZIP file that contains a folder of files you will need for development and testing of Server: phase3.zip
2. The ZIP file contains an extended version of the Simulation class that contains a test driver for the Server class. It also contains the WaitingLine test driver but the call to it has been commented out.
3. I am providing a stubbed-in version of the Server class (click for API). Your solution will build on it.
4. NOTE: Although the main Project 3 page includes the pseudo-code algorithms Simplified Algorithm for Customer Requesting the Server and Simplified Algorithm for Customer Releasing the Server, the code you write for the request() and release() methods will not strictly adhere to those algorithms. The code you write should follow the Server API and the Implementation Notes below. The pseudo-code will come in handy in a later phase when we implement the arrival and departure events.
5. The Server class will require at least two significant instance variables:
   1. a WeightedStatistic object to gather server utilization data, and
   2. a Statistic object to gather service time data
   These variables will be updated within the request() and release() methods.
6. The request() method is used to assign a customer to the server. When this happens, the utilization statistic needs to be collected since the server's busy/idle status changes. The simulation clock time also needs to be put into the customer's timestamp.
7. The release() method is used to release a customer from the server. When this happens, the utilization statistic needs to be collected since the server's busy/idle status is about to change. The service time statistic also needs to be collected since the customer's service is now over. Details below.
8. Service time is calculated by subtracting the customer's timestamp (time it successfully requested the server, obtained from the customer object) from the simulation clock time when it is released. This can obviously only be collected at the time of release.
9. Utilization, collected upon both successful request and release, is more difficult to collect than service time. It is a weighted statistic, therefore every utilization sample also has an associated weight. For utilization, the sample value is 0 if the server is transitioning from idle to busy (successful request), and 1 if the server is transitioning from busy to idle (successful release). The weight here represents time, just as it did for the queue length statistic in Phase 2. To correctly collect the utilization, you need to compute the length of simulation time for which the server was busy or idle. The technique for doing this is similar to the technique you applied in Phase 2 to collect the queue length statistic. .
10. The main() method in the Simulation class contains code to test Server methods. To run the tests, compile and run Simulation.java.

Scoring

To Turn In