This is the eighth 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 relationship between APE and network and application optimization (NAO).
It is worth remembering that one of the major premises of APE is that IT organizations should test applications during the development stage with the goal of proactively eliminating performance bottlenecks. As will be shown in this newsletter, in most cases APE and NAO are complementary solutions.
One of the primary ways that IT organizations implement NAO is by implementing Application Delivery Controllers (ADC). There is no doubt that ADCs provide some functionality that optimizes WAN traffic. For example, ADCs compress the traffic that is transmitted over the Internet. However, no IT organization is going to justify the expense of deploying an ADC merely to compress Internet traffic.
The bulk of the functionality provided by an ADC is focused on improving the performance of server farms. For example, one of the key tasks performed by an ADC is TCP processing whereby the ADC processes the thousands of short-lived TCP connections from the users of the application and hands off to the server farm a small number of long-lived connections. It is highly unlikely that an IT organization that has adopted APE would come to the conclusion that the application could be rewritten in such a way to eliminate the server overhead that is associated with TCP processing. However, as part of the testing and modeling that is inherent in APE, the IT organization may come to the conclusion that it is worth the expense of deploying an ADC to improve the performance of its servers. In this example, APE and ADC are highly complimentary solutions.
Another one of the primary ways that IT organizations implement NAO is by implementing WAN Optimization Controllers (WOC). WOCs provide functionality such as compression and request prediction and spoofing. Assume for the sake of example that as part of the testing and modeling that is inherent in APE, that it is determined that the application as it is designed is very chatty and sends large files.
One option is that the IT organization might decide to rewrite the application to reduce its chattiness and to provide compression as part of the application. Another option is that the IT organization might decide to not rewrite the application but to use WOCs to mitigate the performance bottlenecks. Independent of which option the IT organization takes, the goal of APE was realized – performance bottlenecks were identified during the application development stage and these bottlenecks were eliminated.
The only instances in which APE and WOC are truly competitive approaches is when an IT organization decides that it will not adopt APE to proactively ensure acceptable application performance, but it will deploy an NAO solution and hope that the solution mitigates the performance bottlenecks.
In our next newsletter we will discuss the relationship between APE and Application Performance Management. 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.
