Separate application and network monitoring approaches simply won’t cut it anymore. To be truly successful, your organization must bring these two solutions together.
Recent advances that provide unified and complete visibility across both applications and networks offer the promise of the next generation of end user experience (EUE) monitoring. This combination is the only way to ensure optimal Web performance in today’s world.
Why APM + NPM?
Web-based applications rely on three main components to deliver an acceptable EUE:
1. The Web application itself, which in turn is made up of several layers of software and servers that all have to work together to make the "back-end" work.
2. The end user client/browser, which renders and processes local data coming from the Web application.
3. The mix of data center LAN, Internet and content delivery network WANs, and end-user WAN/LANs, which all have to work together to permit the Web application and end user's browser/client to communicate effectively.
All three of these components must work together seamlessly to deliver a predictable and well-performing application experience to the end user.
The challenge with many application performance monitoring solutions is that they treat the networks as a single "black box." These solutions simply assign one performance metric (most commonly latency) to the network. However, the reality is far more complicated, and solutions must address the effects of loss, congestion, quality of service (QoS), and other factors that dramatically affect the quality of the connection between the application’s server and the end user's browser.
The same is true for the majority of today's network performance management solutions. They're focused on watching device statistics from routers and switches, and they have little-to-no actual understanding or visibility into the transmission control protocol (TCP), HTTP/HTTPs or actual Web transactions happening on the network.
The reality is that networks are built to support applications, and unless both application and network performance are viewed in totality with a full understanding of how different Web applications run on a given network, and how a given network condition affects different Web-based applications, the game of finger-pointing between the network and application teams will continue.
How a Unified APM and NPM Approach Improves End User Experience
Monitoring and measuring end user experience is absolutely the place to start – after all, it all begins and ends with the end user.
If the experience is at or above acceptable business standards for a given application, then everything is good. The challenge is that most EUE solutions stop there. They tell you IF end user experience has fallen below some preset or historical baseline, but not WHY.
Where did the problem start?
What tier in the delivery chain was responsible?
Is the ownership for resolving within direct control of the organization on the hook for delivering end-user experience, or does a third-party need to be contacted to mitigate the issue?
To gain better end user experience visibility, one key is the realization that a single method of measuring performance is simply insufficient. Many solutions get overly hung-up on “their approach being the best approach.” The reality is that it takes a combination of synthetic transactional monitoring, real user monitoring (RUM), and in some cases, even end-point instrumentation to get a total picture of the end user experience. No single approach fits all use cases, so the ability to leverage multiple approaches together offers the most effective path to understand both “what is” (RUM and end point) versus “what is possible” (synthetics).
The bottom line is that if EUE monitoring does not deliver actionable information based on the type of IT professional involved, then all you have is a yet another yapping dog and no way to effectively quiet the noise. To solve the problem, you need an integrated APM and NPM solution that starts with EUE at the top of the stack, and then allows different people within the organization to drill into the details they care about. This will solve the big problem of EUE monitoring and enable the right teams to take fast action to resolve issues when they arise.
What to Look For in a Unified APM and NPM Solution
The key again is to begin with the end user experience - your unified solution needs to start here.
To truly understand the actual end user experience you need multiple methods of measuring, including server-side code injection for RUM, all the way to instrumenting the actual end-users end point (be it a PC or a mobile device). From there, the solution should enable the performance to measure from the true end-to-end of the application service delivery chain.
Once that perspective is obtained, then the solution should be able to understand the details behind a rich, service-oriented architecture (SOA)-type application running over multiple networks and be able to easily and accurately isolate which areas are affecting the performance the most.
As you lead your organization into 2013, it’s important to consider how to keep your performance management strategy up to speed with the growing complexity of your applications and their dependence on the network. Without a holistic approach to monitoring and assuring application and network performance, your team will be stuck in 2012 while everyone else moves forward.
ABOUT Matt Stevens
Matt Stevens is CTO at AppNeta, responsible for guiding technology and product vision and managing the advanced research, development, QA, customer support and information technology teams. Prior to joining AppNeta, Matt was the CTO of the Information and Event Management business unit of RSA, The Security Division of EMC. He joined EMC after the acquisition of Network Intelligence Corp. where he was a founder. In that role, Matt was also part of EMC's Office of the CTO, where he and his peer group had responsibility for EMC's overall strategic technology direction. Prior to NIC and RSA, Matt held senior technology and sales management positions with NetApp, Solbourne Computer and Harris Corporation.