As the IT industry has evolved over recent years, it has seen the clear and rapid movement towards software-defined infrastructure at every layer of the stack. Software-defined, API-accessible infrastructure also opened the door for the next phase of evolution with what Gartner has categorized as “AI Ops”, a further evolution beyond traditional silo-focused, human dependent, operations management.
To create a self-learning system that optimizes Data Center (DC) and cloud resources and predicts and mitigates system failures, I&O leaders should take a holistic approach by combining architectures with smart algorithms, according to a newly released research note from Gartner, “How To Build The Intelligent Data Center1”. I believe this research note explores the approach and vendor ecosystem (which includes Turbonomic) to enable an Intelligent Data Center. We can define an intelligent DC as one that encompasses algorithms, self-optimizing and self- organizing systems (including architecture and applications), operating together to produce an aggregated IQ greater than the sum of its parts.”
Data Center Challenges for I&O Leaders
The challenges highlighted will be very familiar to everyone in infrastructure and operations:
- Infrastructure and operations (I&O) leaders are constantly inundated with hype about artificial intelligence (AI) at the business level, without being offered the necessary guidance for how to raise the overall intelligence level of the data center (DC) and its operations.
- While vendors may sell AI as individual technologies to raise business expediency, there are no standards or guidelines for systems to share intelligence, preventing DC-level insights.
- Rapid rates of change and increasingly complex data centers consisting of decentralized and hybrid cloud infrastructures make it difficult for I&O leaders to select the right tools.
While there is a continuous growth in underlying technologies, there is a lack of products capable of delivering a unified platform to solve the challenges of accelerating the intelligent self-managing data center. And, adding the complexities of a hybrid cloud environment make the challenge even more unsolvable.
This gap results from the lack of the right abstraction across multiple infrastructure layers, which has not been the focus of incumbent vendors whose scope is their own platform. A higher-level abstraction and a holistic data center view is part of a solution set to gain better insights and understanding, which opens the door for an intelligent platform to bring these insights together – across the supply of vendors.
Complex Technical and Business Tradeoffs
There are significant, multi-dimensional challenges plaguing data center environments today. The journey to Digital Transformation will never reach the destination without a fundamental shift that puts this AI Ops vision into reality.
- Improve IT staff productivity – remove the need for human intervention, and automate decision making in day-to-day operations.
- Safely increasing infrastructure utilization – how can we eliminate infrastructure as the cause of performance degradation without overprovisioning?
- Accelerate consolidation and refresh projects – this must be done intelligently, and through the use of AI-driven automation and a common decision engine.
Beyond the virtualized data center, these same practices and platform approach need to be implemented for the public cloud, containers (introducing 10X the complexity of virtual machines on a host), Platform-as-a-Service, and throughout the entire IT stack and lifecycle.
Turbonomic and The Intelligent Data Center
There are some powerfully important pain points in this report, mirroring what we hear from customers. This is true not only in the singular challenges (suboptimal workload runtime placement, unpredictable quality of service, poor compute/storage capacity and utilization efficiency, and others), but also in the collective challenge of managing all of these tradeoffs, at-scale, in real-time.
Turbonomic is a powerful technology that can help IT drive the fundamental shift this report calls for, putting AI Ops into reality. Through full-stack automation, powered by a trustworthy decision-engine, Turbonomic workload automation for hybrid cloud makes workloads “SMART” (self-manage in real-time), in order to assure application performance while increasing infrastructure utilization and efficiency – all while maintaining compliance with business policies and goals.
Learn more about how Turbonomic delivers on the promise of true workload automation for hybrid cloud with a fully-featured trial by visiting Turbonomic.com/download today!
1Gartner Research Note, How to Build the Intelligent Data Center, George J. Weiss, 07 March 2018