RIGHT Blogs                                                              

Application Performance Management

Performance management in general is about optimizing effort and maximizing returns. In the context of software performance management the effort and returns are reflected in the software scalability, speed and stability aspects. Performance issues around response times, downtimes and post production crashes, infrastructure bottlenecks and under/ over utilization of resources impact the ROI on the IT investments. Cost implications apart, such performance failures may damage corporate brand leading to loss of business and revenues. Industry survey results by top research firm indicate that a one second delay in the response time tends to impact customer satisfaction by over 15%.

While enterprises do recognize these issues, the SDLC and QA strategy continue to elude performance testing as it is mostly taken up just before deployment or as a reactive measure to address performance failures. Performance management included earlier into the SDLC will certainly provide better understanding of performance related non-functional requirements which in turn will help eliminate or minimize these issues during the various stages of the SDLC.

By leverage Performance Engineering and Capacity Planning services appropriately across the SDLC, companies can look forward to the following:

1.Better understanding of the application performance characteristics and requirements related to user load, scalability and speed

2.Establishment of performance baseline as per the requirements

3.Performing architectural design review to incorporate performance criteria into the design principles

4.Recommendations on the optimal sizing of the infrastructure based on the scalability requirements

5.Review application code and configurations for optimizing and eliminating application level bottlenecks

6.Testing the scalability of each component at every application layer even before final integration

Apart from having addressed the performance issues before the product moves into the production, companies can also carry out the Performance Monitoring in the post-production stage. At a high-level the services will ensure the following:

1.Monitoring application performance in comparison with performance baseline

2.Performance testing across multiple global locations

3.Identification of the areas of performance degradations

4.Application risk assessment for eliminating application slowdowns and crashes

5.Performance tuning based on product/ applications enhanced requirements for user load and response times

Such proactive application management will highlight the performance limitations of the applications and establishes the confidence levels before deployment. There are various performance metrics that can be captured to assess the performance both from the server-side and client-side. Sample lists of metrics that can be computed include:

1.Memory related metrics – These metrics provide insights on the server-side physical memory in terms of availability and usage, the cache size, page faults, server memory to handle its workload in terms of pages per second etc.

2.Network related metrics – By computing the total bytes per second, one can identify the network activity and the potential network bottleneck by comparing the value obtained against available bandwidth

3.Processor related metrics – These capture the statistics on the processor time spent on processing threads, the processor queue length, the level of interruptions etc.

4.Web service metrics – Web service performance metrics may capture total bytes sent and received by the web server per second, indicate the number of connections to the service during a stress test and other related errors including ‘not found’ errors.

5.Client-side metrics – These metrics will give performance insights related to throughput, response times, hits per second etc. apart from the error counts and error messages arising due to varied reasons.

Performance testing and related metrics definitely enable enterprises to pin point the performance degradations and resolve the issues. The critical aspect here is the timing – handling performance malfunctions before it reaches the customer will be a bigger saver than identifying and rectifying the same after the business is impacted.

The application performance market is definitely growing as there are tools for testing every aspect of performance management and so are there several service providers actively offering these services. With their global delivery models these outsource service providers have the ability to provide performance testing remotely through the performance testing labs across the globe in a low-cost approach. Service differentiation definitely lies in the service providers’ expertise across technologies and their experience in IT consulting to interpret performance issues and provide solution guidance for the same.

To conclude, while almost all industries today offer some transactional facilities to customers by allowing access to applications over the web they definitely run the risk of performance failures. The revenue loss arising from such risks will especially be multifold for industry segments which operate on a business model of providing services to customers over the web viz., financial services, information services, retail and e-commerce. For all such businesses, performance management should be more proactive as it no longer categorizes under a nice to have feature, but is now a must to have non-functional requirement.