QUALITY OF SERVICE AND AOP in Font

Encode ECC200 in Font QUALITY OF SERVICE AND AOP

07 08 09 10 11 12 13 14 15 16 17 18 19 }

public StatsMBeanMixin(Object p) { advised = (Stats)p; } public float getMeanOrderAmount() { if (advised.getOrders() > 0) { return advised.getTotalAmount()/advised.getOrders(); } else { return 0; } }

In JBoss AOP, a mix-in class constructor takes the advised object as a parameter (see line 7 in Listing 9-31). This parameter allows the mix-in class to access the advised object s public members, which here are those of the Stats class (see lines 12 and 13). Finally, we can modify the introduction pointcut in the jboss-aop.xml file, as shown in Listing 9-32. Listing 9-32. The Deployment of the Introduced MeanOrderAmount Attribute <introduction-pointcut class="aop.management.jmx.mixin.Stats"> <mixin> <interfaces>aop.management.jmx.mixin.StatsMBean</interfaces> <class>aop.management.jmx.mixin.StatsMBeanMixin</class> <construction>new aop.management.jmx.mixin.StatsMBeanMixin(this) </construction> </mixin> </introduction-pointcut> To check that the program is operational, this new, aspect-added attribute can be monitored with the JMX program that is shown in Listing 9-33. Listing 9-33. Monitoring the Introduced Attribute GaugeMonitor meanOrderAmountGauge = new GaugeMonitor(); ObjectName meanOrderAmountGaugeName = new ObjectName("OrderProcess","monitor","meanOrderAmountGauge"); server.registerMBean(meanOrderAmountGauge, meanOrderAmountGaugeName); meanOrderAmountGauge.setThresholds(new Float(1000), new Float(500)); meanOrderAmountGauge.setNotifyHigh(true); meanOrderAmountGauge.setNotifyLow(true); meanOrderAmountGauge.setDifferenceMode(false); meanOrderAmountGauge.setObservedObject(new ObjectName("OrderProcess:name=stats")); meanOrderAmountGauge.setObservedAttribute("MeanOrderAmount"); meanOrderAmountGauge.setGranularityPeriod(100L); meanOrderAmountGauge.addNotificationListener( new NotificationListener() {

public void handleNotification(Notification notification,Object handback) { if (notification.getType().equals("jmx.monitor.gauge.low")) { System.out.println("JMX notification - Mean amount < 500 euros"); } else { System.out.println("JMX notification - Mean amount > 1000 euros"); } } }, null, null); meanOrderAmountGauge.start(); The JMX program gives the expected result, which is shown in Listing 9-34. Listing 9-34. The Output of the JMX Example with the Introduced Attribute JMX notification JMX notification Order #1 Order #2 JMX notification Order #3 Order #4 Order #5 JMX notification Order #6 JMX notification Order #7 JMX notification Order #8 Order #9 Order #10 JMX notification JMX notification - Mean amount < 500 euros - Process OK

Summary

In this chapter, you saw how AOP by verifying that an application s definition and execution correspond to its identified requirements can be used to improve the application s Quality of Service. First, we showed how AOP can help to implement contracts in a straightforward way. Second, we showed the AOP implementation of two testing techniques: coverage analysis and nonregression tests. Finally, we showed how to use AOP for the seamless integration of JMX with a Java application. Our AOP implementations did not provide more features than the existing tools do. However, our implementations showed how AOP can be used at a developmental level to improve code quality without needing any specific support. Besides, an important advantage of AOP technologies is that it provides the ability to plug and unplug Quality of Service support when needed even at run time if dynamic AOP tools are used.

his chapter discusses the architecture and design details of a sample application that we ll use as a case study to demonstrate the strengths of AOP within the J2EE environment. This well-known, simple sample application serves to present the material in an understandable manner. The sample is developed from the Duke s Bank application of the Sun ONE J2EE application server. This application does not depend on Sun ONE specifically, and it can be used efficiently on any J2EE application server by adapting the deployment scripts. Since this book focuses on AOP, this chapter does not go into the details of J2EE programming, but instead provides sufficient information on AOP and where it can be useful. A portion of this chapter, however, is a review of the J2EE design patterns used by the sample application. Please note that these patterns are different from the GOF patterns presented in 8. Additionally, we do not present the application s full code here; rather, we simply depict its design. If you d like more detailed information on the application, please visit the Downloads area of the Apress web site (http://www.apress.com) to access the case study code. If you aren t familiar with J2EE, we highly recommend doing some further reading on the subject before you examine the code in detail.