This is the fifth in a series of newsletters that have been discussing a seldom-mentioned IT discipline – Application Performance Engineering (APE). The premise of this series of newsletters is that IT organizations are increasingly concerned with the performance of at least a few of their company’s key applications and because of that concern need to take the proactive steps that constitute APE as a means of ensuring acceptable application performance. In this newsletter we will discuss the linkage between APE and application design and implementation.
This is the fifth in a series of newsletters that have focused on a seldom-mentioned IT discipline – Application Performance Engineering (APE). The premise of this series of newsletters is that IT organizations are increasingly concerned with the performance of at least a few of their company’s key applications and because of that concern need to take the proactive steps that constitute APE as a means of ensuring acceptable application performance. In this newsletter we will discuss the linkage between APE and application design and implementation.
To exemplify one of the linkages between APE and application design and implementation, assume that a global organization is developing a new application. Further assume that as part of the requirements definition phase that the company’s business managers stated that all of the key transactions that are associated with the application have to complete in less than five seconds. As was described in the last newsletter, there are some obvious techniques such as avoiding chatty protocols or implementing admission control that IT organizations can utilize to improve application performance.
However, as was also discussed in the preceding newsletter, WAN delay can have a dramatic impact on application performance. In particular, it is notably easier to meet application performance objectives if all of the users of a given application access the application from a data center that is within their geographic region, be that the Americas; Europe, Middle East and Asia (EMEA) or the Pacific Rim.
The impact of the WAN on application performance brings up an interesting tradeoff that IT needs to present to the company’s business unit managers. For example, assume that the IT organization has done the type of analysis that will be described in future newsletters. Further assume that as a result of this analysis that the IT organization has made two determinations.
One determination is that if the application were only hosted in one data center that the performance objectives would only be achieved in the geographic region that was served by that data center. The second determination is that if the IT organization were to incur additional cost, they could implement the application in three data centers, one in the Americas, one in EMEA and one in the Pac Rim, and with this approach to implementation they would be able to meet the performance objectives.
This information positions the IT organization to go back to the company’s business unit managers and have a fact-based conversation about their previously stated performance objectives. Most likely those objectives were established without any consideration of the cost to the company. Having performed the analysis discussed above the IT organization can present the business unit managers with two alternatives. One alternative does not meet their performance objectives, but is relatively low cost. The other alternative meets their performance objectives but adds significant cost.
In the next newsletter we will discuss how establishing performance objectives impacts application testing. 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.
