Looking Back at 2017 APM Predictions - Did They Come True? Part 1
January 09, 2018

Jonah Kowall
AppDynamics

Share this

I enjoy the end of the year. Getting some downtime from the constant phone calls and meetings allows me to reflect and plan for a new year ... Planning for a new year often includes predicting what’s going to happen. However, we don't often enough look back at the prior year’s predictions to see if they actually came to fruition. That is the purpose of this analysis. I have picked out a few key areas in APMdigest's 2017 Application Performance Management Predictions, and analyzed which predictions actually came true.

Many of the 2017 predictions were not particularly predictive, but instead "observations" of what is happening and how it will accelerate or shift. Without better ground rules for what a prediction is, it's hard to say what the list is and is not. I have picked out a few key areas which encompass several of these predictions. I'm not calling out the individual predictions, which often do not align with Application Performance Management (APM) trends, but instead, serve a specific vendor (gotta love Marketing).

Review the 2017 APM Predictions on APMdigest

The first four key areas below are the predictions that didn’t come true. In addition to the the ones I have called out here, there were many predictions which have been failing for years that vendors wish would come true. Unfortunately for them, that didn't happen in 2017. Hopefully, they don't post the same prediction for 2018. :)

The following few 2017 predictions had several predictions which fell into a category:

The convergence of infrastructure metrics and APM metrics

Although many tools combine APM and infrastructure metrics, most enterprises still use separate tools due to organizational silos

Although many tools combine APM and infrastructure metrics, most enterprises still use separate tools due to organizational silos. The enterprises which run significant infrastructure continue to deploy monitoring tools for each team — meaning teams running data centers still buy tools for DCIM, network, servers, storage, and other infrastructure technologies. Those implementing cloud infrastructure often use cloud provider tools or other infrastructure monitoring tools which do a better job in those environments, and better handle time series. APM tools, have for a while had lightweight capabilities for infrastructure monitoring, and often may replace more capable infrastructure monitoring technologies, especially when migrating to managed data centers or public cloud.

Adoption of platforms to tie monitoring data together

While there are many products which attempt to tie together data coming from multiple monitoring tools, the integration posts challenges for most. Although some of these products are gaining adoption, it's not impactful to the market today. We haven't seen a shift in this market in 2017, and I doubt it will change in 2018.

AI replacing manual analysis

My perspective on AI is documented. AI is still not prominent in solutions today, even though people have been predicting it for years. The advances in making RCA faster have been helpful to users, but the manual analysis of data continues to be standard today, and likely will be for the foreseeable future. We didn't see a shift in 2017, but as I talk about below, predictive analysis and machine learning are becoming more prominent.

Data-centric to event-centric


BUT WILL BECOME TRUE AT SOME POINT

This one was ahead of its time and is a trend which has been occurring for several years. Instead of event-based systems replacing data based systems, I believe there will be an augmented platform to handle both request-response systems and data-based systems. There are many business reasons to have both capabilities within software architectures, and monitoring will naturally have to evolve to handle both types of software systems. Today most of the "event-based" monitoring systems are done with log analysis tools, as tracking events over time is often a challenge for most monitoring tools. There are some which have been able to show change over time, but not managing of event-based architectures. This will be a continuing trend driven by event-based programming and push models for notifications commonly found in technologies built on top of technology such as Node.js or other event-based frameworks.

Read Looking Back at 2017 APM Predictions - Did They Come True? Part 2, outlining the 2017 APM Predictions that came true.

Share this

The Latest

October 18, 2018

Two years ago, Amazon, Comcast, Twitter and Netflix were effectively taken off the Internet for multiple hours by a DDoS attack because they all relied on a single DNS provider. Can it happen again? ...

October 17, 2018

We're seeing artificial intelligence for IT operations or "AIOps" take center stage in the IT industry. If AIOps hasn't been on your horizon yet, look closely and expect it soon. So what can we expect from automation and AIOps as it becomes more commonplace? ...

October 15, 2018

Use of artificial intelligence (AI) in digital commerce is generally considered a success, according to a survey by Gartner, Inc. About 70 percent of digital commerce organizations surveyed report that their AI projects are very or extremely successful ...

October 12, 2018

Most organizations are adopting or considering adopting machine learning due to its benefits, rather than with the intention to cut people’s jobs, according to the Voice of the Enterprise (VoTE): AI & Machine Learning – Adoption, Drivers and Stakeholders 2018 survey conducted by 451 Research ...

October 11, 2018

AI (Artificial Intelligence) and ML (Machine Learning) are the number one strategic enterprise IT investment priority in 2018 (named by 33% of enterprises), taking the top spot from container management (28%), and clearly leaving behind DevOps pipeline automation (13%), according to new EMA research ...

October 09, 2018

Although Windows and Linux were historically viewed as competitors, modern IT advancements have ensured much needed network availability between these ecosystems for redundancy, fault tolerance, and competitive advantage. Software that offers intelligent availability enables the dynamic transfer of data and its processing to the best execution environment for any given purpose. That may be on-premises, in the cloud, in containers, in Windows, or in Linux ...

October 04, 2018

TEKsystems released the results of its 2018 Forecast Reality Check, measuring the current impact of market conditions on IT initiatives, hiring, salaries and skill needs. Here are some key results ...

October 02, 2018

Retailers that have readily adopted digital technologies have experienced a 6% CAGR revenue growth over a 3-year period, while other retailers that have explored digital without a full commitment to broad implementation experienced flat growth over the same period ...

October 01, 2018

As businesses look to capitalize on the benefits offered by the cloud, we've seen the rise of the DevOps practice which, in common with the cloud, offers businesses the advantages of greater agility, speed, quality and efficiency. However, achieving this agility requires end-to-end visibility based on continuous monitoring of the developed applications as part of the software development life cycle ...

September 28, 2018
I developed a Glossary, aimed at introducing topics and indicating where simple further reading can be found about the differences between CS and IT in their applicability to computing needs in today's workplace ...