As part of the Digital Experience Monitoring - Telco's Next Gamechanger' campaign, The Fast Mode spoke to Raen Lim who is the Area Vice President, South Asia at Splunk on how DEM is shaping enterprise IT monitoring as the adoption of Cloud and SaaS becomes increasingly widespread, and as the 'new normal' introduced by the recent pandemic creates new security vulnerabilities and performance implications. The interview is as follows:
Why is there so much interest in DEM?
For many years we've seen growing consumer adoption of online services, and such services were just one of many ways in which companies could reach their customers. With COVID19, online has become the only way to conduct business in many cases, and in such an environment it becomes extremely important to understand and optimize user experience and to quickly detect and resolve any issue. In today's environment any glitch in user experience can result in significant loss of revenue. However, technological developments in web application development, specifically around the increased adoption of Single Page Applications (SPAs) in the front-end and microservices in the backend, mean that older DEM (and APM) tools have significant gaps in visibility leaving the Dev and Ops teams flying blind.
Another trend that we're witnessing is the increasing popularity of OpenTelemetry - the second most popular CNCF project after Kubernetes, supported by Splunk and all the major cloud providers. Previous DEM tools used proprietary instrumentation and proprietary agents to collect telemetry data, effectively locking-in their customers. OpenTelemetry, on the other hand, provides an open standard for enterprises to send their telemetry data. The increased popularity of OpenTelemetry has organizations around the world looking closer at their DEM solution and asking themselves whether they feel comfortable being locked in with one vendor and being dependent on that vendor's roadmap. Because of the increased user demand for online services and limitations of traditional vendors, combined with the rise of OpenTelemetry, you see increased interest in DEM observability solutions that provide visibility into SPAs and full-fidelity tracing to track any front-end transaction to its backend completion.
Raen Lim, Area Vice President, South Asia, Splunk
How does DEM change existing enterprise IT and application performance monitoring?
When looking at online services, in the past, very little code was executed on the user side and the majority of code was executed in the backend applications, therefore if users experienced issues you could reliably say that those issues were caused due to backend problems, and use your traditional application performance monitoring (APM) tools to troubleshoot it. Because applications tended to be monolithic (i.e. a single application running on a single physical or virtual server) that was updated very infrequently the APM tool could just sample the transactions to provide the needed visibility. Additionally, because most of the code was run in the backend, page load times were the logical unit of measure for customer experience.
But now things are different - a lot of code (especially Javascript) and page rendering are executed on the user side, so you need to be able to reliably monitor the performance of that user-side code. Once a page is initially loaded it sends multiple smaller requests for information (known as XHRs) to your back AND to third party services, making the concept of single page load time irrelevant (think of pages with continuous scroll as an example). Because of that, enterprises need to track each and every transaction all the way from the front-end to the backend or to a third party. However, the backend architecture has also changed, and monoliths are broken up into microservices, many of which are updated several times a day. Because of that, APM tools that perform sampling miss a lot of critical information. It's impossible to get a clear picture of the user experience if you need to stitch together a front-end transaction with a corresponding backend trace that was not sampled.
In short, enterprises now need tightly coupled DEM and APM solutions that provide full-fidelity visibility (meaning ingesting, analyzing, and storing all transaction data) all the way from the front-end to the backend with no sampling whatsoever, and with OpenTelemetry providing the standard for instrumentation.
What are your key solutions for DEM?
We recently announced Splunk RUM - the industry's only full-fidelity end-to-end RUM solution. Splunk RUM uniquely combines all front-end spans and transactions with their corresponding backend metrics, traces, and logs in real-time and leverages AI-driven analytics to provide meaningful, actionable insights. Additionally, we have acquired Rigor, providing synthetic monitoring to analyze and optimize web page performance globally before users experience any issues.
These two products, combined with our Observability Suite, which includes best-in-class infrastructure monitoring, APM, logs analysis, and incident response, will help enterprises create the best user experience, detect any issue, and help them troubleshoot it quickly and efficiently.
What are some of the challenges you foresee in getting more enterprises to adopt DEM?
The main challenge is around building a culture and practice that create the best user experience possible. DEM is a tool to help achieve that, but simply buying it won't magically transform poor user experience to amazing one. To really be competitive and create the best experience for your customers you need to make it a strategic priority, and put in place the tools and processes to continuously improve your offerings and have the observability needed to really understand how your offerings are performing and how to solve any problem when it arises.
