Distribution models give you a way to shape control and data traffic rates using a statistical model. Variances in connect and disconnect rates and data packet size can be evenly distributed within a range, or randomly distributed in a bell curve. You choose the range or the width of the curve.
The Normal distribution model allows you to specify a rate or size which is randomly distributed in a bell-shaped fashion. You define the shape of the bell as a percentage of the mean (the base rate or size that you defined) by setting the standard deviation. The width of the bell curve is 4 times the standard deviation.
Example: If you set a connect rate of 200 sessions/second and a standard deviation of 10%, then the actual connect rate will vary in a bell-shaped distribution between 120 sessions/second and 280 sessions/second: 200 - (4 * 200 * .10) and 200 + (4 * 200 * .10) respectively.
The Poisson distribution model allows you to specify a rate or size which is randomly distributed across a Poisson bell curve between a minimum and maximum range. You set the width of this range as a percentage of the mean (the base rate or size that you defined).
Example: If you set a connect rate of 200 sessions/second and a deviation from the mean of 10%, then the actual connect rate will vary in a Poisson shaped distribution between 180 sessions/second and 220 sessions second.
The uniform distribution model evenly distributes the rate or size you specified between a minimum and maximum range. You set the width of this range as a percentage of the mean (the base rate or size that you defined).
Example: If you set a rate of 200 sessions/second and a deviation from mean of 10%, then the actual sessions/second rate will be evenly distributed between 180 sessions/second and 220 sessions/second.
Related Topics