Overview: Web Application Test and Optimization Kitta Original (Participate: 10506, Expert Count: 610) Published: 2003-12-25 1:30 PM Version: 1.0 Read: 3480 times
"Let your web application completed what you want to do is a matter, let them fast, effective to do it is often another thing." In this article, I will initially discuss "Web application" in this article. " Performance "problem, mainly some basic concepts and tools, as bails! Note that these contents are equally applicable to applications such as J2EE. Since then, maybe I will write more articles about this. Of course, it is "perhaps". Ok, I wish us good! First of all, there are two important indicators in terms of performance. Please note that the "definition" below is not standardized for reference only. * Response Time - The time taken from the initialization request to the completion response. This is an important measure for testing the speed of web applications. * Scalability - The response time of a scalable application increases with the increase of the load. No web application can handle unlimited number of requests, but usually we can predict a range and ensure that our web applications can be "elegant" within this range, which always maintains the response time in an acceptable level. If we intend to optimize our web app, then at least we should know that it should not be optimized. Pressure test can answer this question. * Load Test - Pressure Test The process of simulating user requests for web applications to measure their scalability. It is very useful, although it will feel a little "metamorphosis" when starting. Generally we simulate a large number of user requests to obtain the number of concurrent requests that can be processed before the web application is deteriorated to the unacceptable level. The so-called "unacceptable level" is not necessarily to instance pool crash, the application server is paralyzed, and even when the server is machine, this is dependent on demand. The general pressure test contains the following steps: * Determine the maximum allowable delay that accepts the request and completes the response. * Estimate the maximum number of concurrent users of the web application. * Simulate user requests, with a relatively small load, gradually increase the number of simulated users until the corresponding delay of the web application exceeds the maximum delay. * If the load is small than the estimated number of users, this web program should be optimized, otherwise you selectively perform optimization. Would you please consider writing a test program? Forget it, why bother to reinvent a wheel? Here are some pressure test tools, they have features, first introduced: * Web Application Stress Tool, Microsoft, http://www.microsoft.com* jmeter, Java Apache Project, http://www.apache.org / * LoadITUP, Broadgun Software, http://www.broadgun.com If you or your company is very Money, or need to be more rich, you can use the following business software, but they value $ 10,000, or even more: * Webload, Radview Software, http://www.radview.com/; * Silkperformer, Segue Software, http://www.segue.com/; * Benchmark Factory, Quest Software, http://www.benchmarkfactory.com/ * Loadrunner, Mercury Interactive, http://www.mercuryinteractive.com/; No matter which tool you choose, it should at least provide the following features to provide rich and meaningful test data: * Send GET and POST Request * "Record" The GET and POST requests sent from the browser (so as to prevent developers need to write these suitable requests). * Get and send cookies.
