Skip to main content

The link between AIOps, DevOps and CloudOps

(Image credit: Image Credit: Sergey Nivens / Shutterstock)

The artificial intelligence hype machine is on overdrive… again. Allied Market Research predicts that the market for Artificial Intelligence as a Service will skyrocket from $2.39 billion in total revenue in 2017 to $77.04 billion in 2025. That represents a whopping compound annual growth rate of 56.7 per cent. Business Insider offers an overview of the study findings.

(Personally, I’m waiting for the Beer and Chips by the Pool as a Service market to take off, but as with AI, BCPaaS is going to take more than wishful thinking.)

Some say we have entered the first of the three stages of AI. Many current product purporting to be AI-based use ‘narrow’ AI, which focuses on a single task and operates in a limited context. This encompasses weather prediction, sales forecasts, purchase suggestions, and similar uses.

By contrast, artificial ‘general’ intelligence is intended to solve novel, unpredictable problems by applying reason and context, as a human would. Lastly, artificial ‘super’ intelligence actually surpasses human cognitive capabilities, which is fodder for utopians and dystopians alike (i.e. Skynet).

One of the hot buzzwords of 2018 and continuing into 2019 is AIOps… using those narrowly scoped use cases to help IT operate smarter.  There are multi-cloud management platforms out there, whose intelligent analytics applies machine learning to gain fresh insight into how your VMs, containers, and public clouds are being used and more importantly what the under-utilised instances are costing you. These remediations let you resize app components, set power schedules, and collect information on brownfield installations.

AI’s rosy future depends on cloud services — and a big dose of reality

Without the availability of cloud data and services, much of the promise of AI would never translate into business success.  This is because many of the large data sets that are the foundation of analytics would not be accessible without cloud access.  Secondly, the scalability and cost efficiency required to run data-intensive apps affordably are impossible to furnish without the public cloud.  Lastly, cloud native services and functions are emerging which simplify and accelerate the development of practical analytics applications.

In addition to accessibility and affordability, cloud-based AI addresses the shortage of skilled AI practitioners to create the applications.  One of the main drivers of multi-cloud adoption is the fact that different cloud providers have different native services, some of which are better suited for different types of AI applications.  It’s one reason multi-cloud management has heated up as it allows IT to align projects to the best execution venue.  Even for on-premises projects there is a requirement to rapidly deploy new databases and application stacks.  This also plays to the strengths of DevOps-centric multi-cloud management platforms. Want a 7-node big data cluster with a seed data set?  Hit the easy button.  Need to re-deploy that same cluster into a different cloud, say for data gravity or data sovereignty reasons?  Same easy button… different cloud target.

So what’s holding up adoption of AI-based cloud services? To start with, there are the challenges of securing and maintaining regulatory compliance for the massive amounts of data AI apps require. Equally daunting for many companies is the complexity entailed in developing, deploying, and maintaining AI applications. As a result, firms are choosing to partner with cloud services that give business managers access to advanced analytics tools without the need for extensive training in how to put them to use.

Capture more value in less time to make better decisions

Everybody knows data is valuable, but not many companies know how the value of their data affects the bottom line. A McKinsey Global Institute briefing states that the growth in the volume of data collected by organisations hasn’t been matched by commensurate growth in revenue and profit. Many companies are applied advanced analytics to gain a competitive advantage through faster evidence-based decision making, insight generation, and process optimisation.

The McKinsey report outlines the components of effective analytics transformations:

  • Match your business strategy to the apps and data sets to generate insights that maximise value.
  • Optimise your data-collection processes–particularly those associated with legacy systems–to capture more value from customer interactions, suppliers, and internal procedures.
  • Either hire staff with the skills required to apply advanced analytics to your data, or partner with a specialist that puts powerful analytics tools in the hands of DevOps and business managers.
  • Incorporate the insights you gain from analytics in your everyday workflows and business processes to ensure decision makers have access to the highest quality information.

What sets cloud management platforms apart as drivers of advanced analytics is the ability to create new services quickly from existing underlying core services, as Forbes Technology Council member Kristof Kloeckner writes. New AI services are available first on the cloud, which heightens the cloud’s influence on standards. Standardisation enables the automation that is a prerequisite for manageability at scale and “industrial strength service delivery.”

Another benefit of cloud standardisation and automation is “DevOps for AI,” which Kloeckner believes will result in accelerated service delivery life cycles and faster business cycles. If data is the raw material of the 21st century, AI is the refinery, and optimised business processes are the end product. This self-reinforcing chain will accelerate to keep pace with changing markets only if data science and DevOps skills are available, and your company’s stakeholders agree on a clear strategy that identifies intended benefits and management risks.

Brad Parks, Morpheus Data
Image Credit: Sergey Nivens / Shutterstock