We don’t just find performance problems, we fix them too

The point of performance testing is not just finding but fixing performance issues before go-live. Lots of people can drive load, take an x-rays and hand them to you. We don’t leave you wondering “Now what do I do”? We take you to the next level: we interpret the results, diagnose the causes, and provide the cure.

Our Approach Before conducting the test, we guide you in configuring monitors on each of your system components. Using a combination of operating system and specialized monitoring tools, we measure the bellwether metrics of servers, system software, storage and network components.

During the test, our deep-dive database, middleware and network experts watch these monitors as load is ramped up for tell-tale signs of performance degradation and limiting resources to provide you initial clues on potential causes. After the test, we correlate these metrics against response times and load to hone in on likely causes. We then work with your technical team to fully diagnose the issues and help you apply the fixes. We then re-test to validate the effectiveness of the fix..

Components we have expertise in

• Databases
  From experience we know that 50% of the time the problem is in the database. We use specialized tools for Oracle and SQL Server to monitor locking contention, access methods, buffer management and SQL compilations to identify the parameters that need to be tuned or queries that need optimization.
• Application Servers
  Often the problem is in the application services middleware. Whether your system uses WebLogic, WebSphere, Tomcat, JBOSS, Blue Martini, MS Commerce, JMS or MQ messaging servers, we know how to monitor heap memory, db connection pools, and message queues to help root out the problem and tune the appropriate configuration parameter.
• Storage
  Sometimes the problem is IO bottlenecks. We’ll show you how to monitor and correlate IOPS on SANs, NAS and other storage arrays.
• Network
  If the problem is in the network, we’ll use monitors and a network sniffer to reconstruct the tcp and http conversations to localize the problem, whether on your network segments, firewalls, load balancer, network interfaces, or Citrix infrastructure, and guide you in resolving it.