IT operations (ITOps) has always been rooted in data gathering and analysis. Now artificial intelligence (AI) and machine learning (ML) are being applied to enable a new class of Ops tools that can actually learn and improve from the data they gather. The advancements haven’t come a moment too soon, as the IT crisis created by the COVID-19 pandemic has forced organizations to stand up widely distributed applications and infrastructure. The emerging class of ITOps tools being called AIOps promises a solution to the sudden complexity.
David Linthicum, in his recently published report “Best Practices in Moving from ITOps to AIOps,” explores the journey that IT organizations face as they seek to leverage ML and autonomous system interaction to speed diagnosis, reduce downtime, optimize infrastructure, and anticipate challenges.
Linthicum breaks this journey into four stages: ITOps, Emerging AIOps, Advanced AIOps, and Future AIOps. The progression starts with a traditional approach built around siloed IT monitoring, script automation, and manual Ops processes, and ends with process workflow, predictive automation, and business automation.
“Note that we move from restrictions of the traditional approach to an emerging use of AI Operations,” Linthicum writes in the report, citing the infographic in Figure 1. “This has a few core attributes, such as the ability to monitor systems using correlated data, the automation of runbooks, the ability for the AI engine to learn from data, and the ability for much of this functionality to be provided on-demand, as needed by the Ops teams.”
Figure 1: Phases of AIOps Adoption
Ultimately, the goal is to adopt concepts of autonomic computing, which refers to the self-managing attributes of distributed computing resources and their ability to adapt to unpredictable changes while hiding complexity from both operators and users. In other words, as Linthicum notes, “the ability to remove the humans from the underlying operational complexity.”
In the report, Linthicum offers best practices to help IT organizations embark on an AIOps journey. The guidance starts with issues of planning and measurement—considering the business problem, mapping the course to achieve AIOps, and getting a handle on the value to be enabled. From there, he explores movement: Transitioning into advanced concepts like predictive analytics and self-provisioning while implementing a continuous improvement process for AIOps and ensuring integration with other Ops tools. Finally, he urges introspection, evaluating value and grading performance, while also applying a continuous cycle to the ongoing AIOps effort.
As Linthicum observes, the use of operations automation tooling is a “forgone conclusion,” but that doesn’t mean it will arrive in time to address the skyrocketing complexity of IT infrastructures. He urges organizations to map out their AIOps journey early as a way to prevent getting surprised.
Learn More: Best Practices in Moving from ITOps to AIOps