Apply a statistical distribution model to a rate or packet size to simulate real-world activity by varying control traffic rates, and data traffic rate and packet size. If you select Fixed, the rate or size that you specified in the test definition is used throughout the test. If you select a distribution model, you can further refine the model by defining or limiting the deviation from mean.
Options: Fixed, Uniform..., Normal..., Poisson...
Default: Fixed
One of these parameters may be available depending on the model selected.
Standard Deviation — defines the standard deviation as a percent of mean in the Normal distribution model.
Range: 0-100
Default: 0
Maximum Deviation from Mean — limits the standard deviation from mean as a percent of mean in the Uniform and Poisson distribution models.
Range: 1 — 99
Default: 10
Distribution models can be used to vary the following rates and volumes:
Mobility Rate Interval Distribution is used in the handoff and handoff rate test activities. It applies a distribution model to the handoff rate.
The following measurements are affected by Mobility Rate Interval Distribution:
Packet Size Distribution is used in the Data Message Flow window. When you use one of the Basic Data protocols, you can vary the size of the messages sent by the message flow by applying a distribution model to the data payload Packet Size.
This distribution model affects the number of bytes sent by the MN sessions and the number of bytes received by the Network Host in a given interval:
Revocation Rate Interval Distribution is used in the MIP Revocation test activity. It applies a distribution model to the Revocation Rate.
The following measurements are affected by Revocation Rate Interval Distribution:
Transaction Interval Distribution is used in the Data Message Flow window. It applies a distribution model to the Transaction Rate used in a message flow.
|
IMPORTANT: When a distribution model is applied to the Transaction Rate, the additional processing required to manage every MN data stream will greatly decrease the maximum throughput achievable. Results will vary depending on the model selected and the deviation or range defined. These models should not be used for performance testing. |
This distribution model affects the throughput measurements for a given interval: