Virtualization Technology News and Information
Zenoss 2019 Predictions: No Skynet here...

Industry executives and experts share their predictions for 2019.  Read them in this 11th annual series exclusive.

Contributed by Mike Lunt, VP Engineering, Zenoss

No Skynet here...

20-ish years ago, Agile was acquiring mainstream pickup, and now, triggered by the Agile movement along with some new enabling technologies, DevOps is doing the same with a decade's worth of momentum. While transforming into a DevOps culture is practically a do-or-die imperative for most businesses, the incorporation of artificial intelligence (AI) into business-generating ideas is at the beginning of its hype cycle. Since many of last year's predictions about AI turned out to be a bit premature, 2019's forecast is focused more on the enabling steps necessary to enact a bona fide world of machine-savviness.

Infrastructure becomes completely blurred

Physical and virtualized infrastructure vendors originally tackled the inevitable conquest of public cloud by providing features that allowed on-prem resources to extend/burst into public cloud. As with most hype cycles, the promise of fully transparent infrastructure has yet to become a reality, but a few notable vendors are setting the tone for what's to come. Google's GKE On-Prem reminds us that good ole on-prem infrastructure still has a purpose, and on the opposite side of the tracks, Nutanix's Prism and Beam remind us that enterprise customers are demanding a seamless management experience across both on-prem and public cloud infrastructure. While the operational complexity of managing this matrix of multi-cloud and hyper-converged infrastrastructure will only increase, the promise of write-once-run-literally-anywhere is finally on the visible horizon.

Context and performance become overused

As any enterprise IT veteran knows, industry analysts provide a never ending normalization of jargon-y terms and acronyms as a way to make comparisons and help explain a highly technical world. Service health and service assurance have been the business driving goals for IT operations management (ITOM) tools for a few decades, but given the DevOps disruption of recent years, performance monitoring of the entire IT stack has become the new battleground given the use of highly ephemeral software nuggets (microservices) strewn across semi-invisible infrastructure. Application performance monitoring (APM) vendors initially seemed to have an eye towards the future by promising end-to-end visibility from the user through the application server; however, DevOps teams need context of the full IT stack and don't want to wait for a traditional NOC operators to perform calibration of the monitoring tool(s). Immediate and self-served performance monitoring of the entire IT context will replace legacy terms such as service and visibility to become the go-to qualities that attract DevOps cultures with freshly formed Site Reliability Engineering (SRE) teams.

IT ops realizes its tools are not intelligent

The "unified" part of unified management tools continues its evolution from breadth+proactive to breadth+automation to breadth+self-heal. The "breadth" part of this equation has broadened from being focused on a range of monitoring targets to adding a range of platform capabilities; and most recently, machine learning (ML) with a lean towards AI has become trendy. As ML-touted tools become status quo, operations teams are already learning their existing ITOM tools need expansive breadth to create meaningful ML models. This means that adding intelligence on top of logs or events alone is not sufficient to create trustworthy automation. For operations teams to realize robust self-healing, tools must be able to stream varying types of data from all sources to provide the enough context to be trustworthy in hands-off scenarios. Ops teams are beginning their journey to cloud-based tools with the scalability to combine events, model, metrics, logs and etc., and only tools with the ability to intelligently measure performance from all forms of collection will become viable mechanisms for autonomous data centers to flourish.


About the Author

Mike Lunt 

Mike Lunt is Zenoss's vice president of Engineering. Mike oversees Zenoss's global engineering teams, which are responsible for both the open source and commercial products at the heart of Zenoss's offerings. Mike joined Zenoss in 2008 and has over 20 years of experience delivering enterprise on-prem and SaaS based solutions to the Fortune 500. Prior to Zenoss, Mike served as a director of R&D at BMC Software, responsible for delivering various ITOM solutions, and he was a key change agent in BMC's transformation to Agile development techniques. Mike joined BMC through the successful acquisition of Evity, providing Web transaction monitoring solutions, where he lead the Operations and QA teams. Prior to Evity, Mike was a founding engineer of, which was acquired by, and a part of other early stage companies such as Aquity. Mike holds a Bachelor of Science in Mechanical Engineering from Oklahoma State University. 
Published Thursday, October 25, 2018 7:53 AM by David Marshall
2018 Second Half Publications – Continual Improvement - (Author's Link) - January 13, 2019 3:37 PM
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<October 2018>