skip to Main Content

What Is Self-Driving IT Operations?

What Is Self-Driving IT Operations?

According to the second law of thermodynamics, the entropy of an isolated system never decreases. In computer science we’ve traditionally described entropy as a measure of the disorder and randomness of a system. Taking some liberties, then, we might say that, in general, we expect the disorder and randomness of any system to either stay the same or increase over time. That’s exactly what has been happening with our data centers for years now, and it has recently begun to reach a critical mass where our ability to scale our management and monitoring oversight to match is being woefully outpaced by the increase in complexity.

What Makes vRealize Operations Self-Driving?

The rapid increase in the scale of our data centers, from the density of storage and virtual machines (VMs) to the speeds of the network presents a challenge. And the monitoring and managing of workloads, events, triggers, and movements of those elements even more so. We can no longer effectively oversee everything using the same tools that have served us well in the past. The last generation of tools relied entirely on humans responding to static dashboards, or to triggered events which we configured ahead of time—manual human responses to things we already expected to happen. When we had smaller numbers of applications, workloads, and storage pools it was workable, now it’s all but impossible without better tools. That’s where VMware vRealize Operations comes into play.

Just as cars are becoming self-driving by utilizing components of artificial intelligence (AI), vRealize Operations uses similar advancements to increase the intelligence of data centers. As those data centers become self-driving, utilizing machine learning and deep learning to provide continuous, automated, closed-loop performance optimization, more efficient resource management, and densification and proactive planning, the operational efficiency and organizational IQ of your business increases.

What Is Closed-Loop Performance Optimization?

A key component of achieving all of this is closed-loop performance optimization. According to Gartner, closed-loop performance management (PM) is “the discipline of “taking action” on the results of performance monitoring, and blending these new modified results with updated plans and goals to drive business value and impact”. But before we can fully define that we must understand the concepts of operational and business intent. Simply put, both are abstractions we apply to our technology decisions in order to help us as IT practitioners contribute to business needs and wants. If a business segment owner comes to us and describes the use case for a particular application, and the application owner comes to us and describes the operational requirements of the application, it’s up to us to decide how to place that application in our data centers—on what storage tier, across what size and number of VMs, what level of redundancy and backup policies, what network speeds, and so on. Business intent is what the business wants, and operational intent is how we make that happen within the context of IT. That used to be a very manual process, applied to each and every application. With virtualization, containerization, micro-services architectures, and any number of other advancements, manually managing those decisions has become untenable.

Why Adjusting to the Unknown Is Key

With the combination of vRealize Operations and vRealize Automation, however, those decisions can be largely managed by our self-driving data center. Initial and ongoing placement of workloads across clusters can be based on business and operational intent, designed from the inception of the data center. Availability, compliance, utilization, and license costs can all be baked into our system, and our data center can now make intelligent decisions as to where workloads not only should go initially, but what should happen to them as the environment changes. What if a data center loses connectivity or is degraded? What if a storage tier is becoming full or has another problem? The system has the ability to react in real time to changing conditions and can move workloads, change storage tiers, increase or decrease scale-out performance, all with no human involvement. We are still in full control of our data centers, however, and can take the wheel if we need to for troubleshooting or manual intervention of any kind. Self-driving IT operations doesn’t mean we lose control, it means we gain control, insight, and freedom, becoming more efficient and valuable to the business.

The complexity of data centers will continue to increase exponentially, and the only way we’ll keep up with this rapid pace of change is by leveraging our ability to create more intelligent systems, systems that can increasingly use the tenants of AI to make appropriate decisions as quickly as the data is changing. With its deep history and pioneering work in the data center, VMware is in a unique position to leverage machine learning to positively impact the way its customers manage and operationalize their infrastructure and workloads, and to usher in a new era of data center operational efficiency. vRealize Operations is a key component of that strategy.

For more information, check out the blog post, “Under the Hood: What’s new in Self-Driving Operations with vRealize Operations Manager 6.7.”

** This post was originally published on **


×Close search