This is the seventh in a series of newsletters that have been discussing a seldom-mentioned IT discipline – Application Performance Engineering (APE). In this newsletter we will discuss the limits of tools that are based entirely on mathematical modeling.
As mentioned in the previous newsletter, application development organizations need to be able to test application performance throughout the application life cycle. As was also mentioned, in most cases today the tools they use to do this testing are based on the use of mathematical models. For example, the tool might assume that the traffic volume on the WAN follows a bell curve (sometimes call a Normal or Gaussian Distribution) or some other well known distribution. Sticking with the example of making the assumption that the WAN traffic follows a bell curve, the determination is then made that the distribution of WAN traffic volumes has a mean of 650Kbps and a standard deviation of 100Kbps. The advantage of this approach is that there is a well-known equation that defines a bell curve.
Tool developers can use this equation to develop tools that simulate the performance of a WAN on the assumption that the WAN traffic volume follows a bell curve. The disadvantage of this approach is that it is an approximation. The real production network does not follow a bell curve or any other simple mathematical model. Hence, while it might seem as if tools based on mathematical models are highly sophisticated, the reality is that the accuracy of these tools is limited by the accuracy of the assumptions that are used to create the mathematical models.
That does not say that the use of tools based on mathematical models is not sometimes valuable. It is. However, where possible, application development organizations should avoid the limitations of tools based on mathematical models and should utilize tools that capture and utilize the actual performance of the production network.
In our next newsletter we will discuss the tradeoff between APE and network and application optimization. In the meantime we are interested in your experiences with any form of APE. Kindly write to us and let us know what you did, what worked and what didn’t work.
