Today, enterprises rely heavily on applications for nearly all business-critical processes. These applications are delivered via a combination of hardware, software and services, known as the application delivery chain. To make this delivery chain work together effectively, IT must take a new and converged approach to network performance monitoring and application performance management.
Network and application performance issues are growing dramatically in importance for a variety of reasons including complexity of data centres, number of devices and rising end-user expectations. A survey conducted by Compuware earlier this year found that 60 per cent of companies in Australia identify network problems more than five times a week. The same survey also found that sixty-three per cent of companies believe that cloud-based services has impacted troubleshooting network performance.
Today, IT organisations need a detailed, quantitative understanding of whether their applications effectively meet their business objectives.
Base-level Network Performance Monitoring (NPM) that delivers statistics concerning the protocols and infrastructure that support business-critical applications is certainly helpful. But it cannot produce all of the metrics pertinent to application performance and understanding the end-user experience.
In the same way, an application-centric approach to monitoring performance (or APM) is incomplete without insight into the underlying architecture and protocols that ensure successful delivery of the application and that provide the ability to isolate performance problems from an end-to-end perspective.
The limitations of bridging solutions
To fill this gap, vendors from both ends of the spectrum are developing new NPM and Network-Aware APM solutions, yet neither approach as yet has managed to fully correlate data from the network and applications to the expectations of IT and Operations managers.
Due to the difficulty of collecting information and using it to provide context to the application, NPM is often only effective in troubleshooting, unable to provide real-time analysis or true application performance monitoring, in relation specifically to the end user perspective
On the other hand, Network-aware APM tries to provide insight into the supporting infrastructure, but is too heavily focused on application-layer transaction analysis and has limited visibility into the network. The resulting lack of combined insight tends to lead to either a vague determination or misdiagnose of the ultimate root cause. In either case, the result is extended problem resolution timeframes and a view into only a portion of the application delivery chain.
The convergence alternative
A more holistic and therefore successful approach is the convergence of NPM and APM solutions, ultimately producing what is best described as a new generation of UEM and aaNPM solution
Network performance has a direct impact on application performance and it is critical to providing an accurate picture of performance across the entire application delivery chain. An integrated UEM and aaNPM solution carefully correlates metrics at the network level with application performance management data in real time.
It provides IT and operations managers with an in-depth understanding of the underlying technologies that support applications, enabling them to make more effective decisions regarding IT investments, problem resolution, and resource utilisation and allocation.
Meeting service level agreements
One activity where a converged UEM and aaNPM solution can help is when monitoring performance against service level agreements (SLAs). To confirm whether applications are meeting SLAs, quantifiable data can be derived by correlating NPM-related metrics with application transition details.
The first step is to obtain baseline data about application delivery. This baseline will be something of a moving target and must be regularly updated with production data to ensure it adapts to the enterprise environment as hardware is updated, new versions of operating software are rolled out, or as additional users are added. However, once established, the baseline allows easy identification of any future irregularities in behaviour or performance degradation.
An up-to-date baseline also provides IT and operations with the ability to more readily identify system potential, proactively identify problems, improve the troubleshooting environment, expedite mean time to repair (MTTR) and make informed decisions about IT investments, resource utilisation and allocation.
Another benefit from converging APM and NPM is the ability to accelerate troubleshooting. Fast identification and resolution of issues requires accurate performance information from not only the application but also from the supported infrastructure and protocols.
This is one reason why silo-oriented monitoring tools have proven to be ineffective early warning solutions for application service; they don't measure the end-user response time or provide any cross-domain intelligence. Network information must be correlated with application specific data for targeted troubleshooting and to enable isolation of the fault domain – whether it’s the server, application or network – all while understanding the business impact.
Knowing what's possible
A converged NPM/APM solution arms IT and operations managers with invaluable knowledge about what is possible given their existing system. This enables them to more readily respond to commercial changes or implement new technologies that can enhance business practices.
When performance management is viewed from an application perspective, end-to-end through the entire infrastructure including client devices routing protocols, network configurations, servers and associated components, it creates transparency and a common understanding of performance degradation.
Network performance monitoring is a necessary discipline in the context of an APM solution. On its own, it can produce important metrics for reporting, but to contribute to the overall delivery story it must be combined, viewed and reported in conjunction with the other elements and data provided by an APM solution.
Just as the performance of a network and the applications it supports are deeply intertwined, it is necessary to have a truly converged Network Performance Monitoring and Application Performance Management approach for a holistic understanding of end-user experience.
Mike Hicks, Senior Product Manager, Network Performance Management, Compuware