Businesses want applications to work well, so why not get user feedback? A case for real-time situation awareness in application performance monitoring.
User Experience Monitoring (UXM) is an effective approach for understanding client-side activities to provide the foundation for analysis, detection, and resolution of performance issues.
UXM is today a distinct yet important aspect of Application Performance Management. The combination of UXM and APM provides the ability to monitor all aspects of how an application is performing on the client, through the network, and on the server. There is no question that having a complete view of how the application is performing is indispensable in ensuring that applications stay as true as possible to their stated or intended service-level agreements.
Real-time data difficulties with APM
The foundation underpinning APM for keeping applications on track is nearly continuous fine-grained data collection based on applications, events, sessions, transactions, and packets traversing the network. There is little doubt that tracking every click, change of state, and transaction can provide a wealth of knowledge regarding how an application is being used and how effectively it is performing. However, there is significant complexity in tagging this data (time stamps and identification data), evaluating it against policy in real-time or near real-time, and then presenting this data in proactive or reactive form in support of process remediation or improvement efforts. This may explain why APM is not routinely applied to every enterprise application.
Riverbed’s recent acquisition of Aternity provides them with useful UXM capabilities that will soon be well integrated into the SteelCentral portfolio. Riverbed was quick to point out that Aternity provides them with a way to monitor client-side activities and this client-side information alone is sufficient to identify performance concerns. This means that client-side monitoring can be used to triage performance issues and then bring the comprehensive capabilities of SteelCentral to bear where needed. This seems like a pragmatic approach that is fully data-driven and transparent to the user of client-side device. However, it still requires a mountain of real-time continuous data collection and analysis to make this client-side approach work.
Getting feedback from end users
What about involving the end user in this process? While you think about this, keep in mind that UXM activities only really reside in a B2B setting where users are someone’s employees. This means that users are likely to care about the performance of those enterprise applications they are running on a device because they make the job they do possible or easier. Therefore, these users are at least partially invested in having applications that work well when they need them. Therefore, why not allow or recommend that the user participate in providing feedback on an application’s performance?
This concept has been pioneered by Phillip Hasse in what he calls Integrated Real-time Situational Awareness (IRSA). If users begin to experience issues using an application, they would simply click a button much like today’s thumbs up/down button that you see in consumer applications like Pandora, Facebook, or Linked-In. This event could then propagate a variety of actions such as a snapshot of the application configuration, initiate a period of continuous monitoring, logging a trouble ticket, or launching a management application.
Advantages of end-user experience monitoring for APM
An obvious advantage of this approach is that it doesn’t rely on continuous real-time data collection or analysis across a vast sea of employee devices. Another advantage of this approach is that it links application performance to a specific dissatisfaction instance on the part of a user, and in the context of when, where, what, and how the application was being used.
This kind of context is invaluable in helping reduce the mean-time to resolution (MTTR) for incidents. Another advantage of this approach is that it steers clear of privacy concerns that are bound to be voiced by employees in BYOD environments or employees who simply find this level of intense enterprise oversight too intrusive. A disadvantage of this approach is that is likely to be less effective at proactively identifying developing problems when used without data-centric UXM solutions.
What’s unusual is that IRSA is not a capability that has made its way into the lexicon of APM activities. Perhaps this is due to the patent (#9106517) that Mr. Hasse has on the IRSA approach. Or perhaps APM vendors simply haven’t pursued UXM far enough to think about this kind of a capability. What’s a given is that IRSA-like capabilities are a natural fit for any comprehensive end-to-end APM or network performance management solutions.
