Reference applications come in all shapes and sizes. They're generally designed to help developers learn how to use new languages, APIs, or frameworks, but they can also be extremely useful for evaluating application performance management tools, such as Riverbed® SteelCentral™ AppInternals.  In this blog post, I'm going to share my experience using the AppInternals component of SteelCentral to evaluate end-user experience with a reference application called TicketMonster.

 

TicketMonster is a modern web application that is representative of an online ticketing broker - providing access to events (e.g. concerts, shows, etc) through an online booking service. This application uses both Java and .NET web services in a multi-tier heterogeneous environment that is typical for many AppInternals deployments. The application map looks like this:


application map.png

 

The load for this application was produced in a fashion that simulated a global user base using all kinds of browsers and operating systems. AppInternals reflected these usage patterns nicely with visualizations such as those shown below:


geo.png

users.png


We can use this environment to exhibit a realistic workflow for diagnosing delays which occur when customers attempt to purchase large ticket orders. In a business context, this is exactly the type of problem that would have urgency in the real-world, since users placing large orders could conceivably be the "power users" who generate the most revenue on our site.

 

The following video demonstrates the powerful search capabilities in AppInternals to make three key observations:

  1. Determine how often users experience slow ticket orders
  2. Identify a root cause (slowest class and slowest method)
  3. Identify the email addresses for affected users

 

 



After seeing this demo, the uninitiated may be inclined to ask,

  • "I can trace a transaction even as it executes across different servers?"
  • "I can trace a transaction that happened weeks ago?"
  • "And I even have access to call stacks that illustrate execution paths through both Java and .NET services???  WOW!"
  • "The capability to instantly access so many details for transactions in my production environment is something I absolutely must have!  Where have you been all my life!"

 

stick figure dialog 02.png

 

 

The TicketMonster application is well suited for showcasing many of the capabilities of AppInternals. We were able to quickly pinpoint and diagnose performance problems in TicketMonster and get immediate answers to complex business problems, such as "How can I identify our most valuable users and detect when they experience poor application performance?".


For more information about application performance management with SteelCentral AppInternals, visit www.riverbed.com/products.