We are now living in a digital economy. Organizations all over the globe are turning their much of their focus away from producing physical assets, and towards designing and developing digital products and services that either complement or completely replace their physical predecessors. Digital transformation initiatives are at the very top of business agendas across all industries, resulting in big changes and big demands being placed on enterprise IT operations.
This shift isn’t coming out of the blue, of course. For many years, high demands have been placed on IT operations teams – and DevOps in particular – to become more agile and proactive so that their businesses can quickly embrace new technologies and practices in order to remain competitive. However, to meet these demands, IT operations has faced the challenge of keeping costs down on the one hand, while dealing with the increasing complexity of operations on the other. As such, it should come as no surprise that automation and machine learning has found its way into the IT operations model as a means to tackle both issues – and it’s beginning to look like these technologies will be the core driving force shaping the future of IT ops over the coming years.
Coined by Gartner, AIOps – or Artificial Intelligence for IT Operations – is a solution that uses machine learning-powered algorithms to solve known IT issues and intelligently automate repetitive tasks previously carried out manually by IT staff. In essence, AIOps is about reducing the amount of work for IT personnel when dealing with the vast quantities of machine data generated by infrastructure hardware and software.
AIOps platforms are capable of deriving meaning from large data sets on their own, with or without human input. What’s more, AIOps denotes a shift away from siloed IT data. Instead, machine data from multiples sources is combined under one platform for analysis, resulting in intelligent insights that can be used to both predict possible incidents, and trace back in time to determine the root causes of current system behavior.
See our previous post – ‘What Is AIOps?’
Why AIOps Is the Future of IT Operations
The main flaw of current system monitoring tools – many of which were built five or even ten years ago – is that they were not built to meet today’s big data demands. These demands are defined by the “three Vs” – volume, variety, and velocity. With a two- to –three-fold increase per annum, the rapid growth of data volumes generated by IT infrastructure and applications is enough on its own to stretch current monitoring tools to breaking point. But, add into the mix the increasing variety of data types generated by machines and humans (metrics, logs, wire data, etc.) and the increasing velocity at which data is generated, and it becomes almost impossible for these outdated tools to extract the fourth and most important “V” from big data – value.
Of course, processing all incoming machine-generated data at high speed is more than human beings are capable of – but is precisely what machine learning algorithms excel at. Artificial intelligence is designed to do what humans do, only better, faster, and at scale – and the promise of AIOps platforms is that they will address the speed and complexity challenges of digital transformation in IT operations.
So, what are the driving forces behind AIOps? Well, first, it’s the difficulties IT operations currently face in manually managing infrastructure. Modern IT environments are far more expansive and scattered than they used to be, and today include both managed and unmanaged cloud, SaaS integrations, third party services, applications, mobile, and much more besides. Managing this complexity is already beyond the scope of current IT operations technology, and environments are only set to become even more complex over the coming years.
Then there’s the sheer amount of machine data IT operations has to deal with. Performance monitoring across the environment is generating huge numbers of events, alerts and service tickets – and again, it is becoming too complex for manual reporting and analysis. Next, we have infrastructure problems, which must be responded to at ever-increasing speeds, as negative impacts on end user experience must be kept to a minimum.
It’s IT’s job to keep the lights on and provide stability for the infrastructure that organizational applications rely on. But as organizations increasingly digitize their business, IT becomes the business. As such, – and as Gartner puts it – IT operations need tools that can help:
- Extrapolate future events to prevent potential breakdowns
- Initiate action to resolve problems proactively
- Capture anomalies that go beyond static thresholds to proactively detect abnormal conditions
- Provide better causality, which helps identify probable cause of incidents
- Reduce noise (in the form of false alarms or redundant events)
AIOps platforms are designed to deliver on all these fronts, and, as a result, greater adoption of AIOps strategies and technologies will almost certainly manifest across enterprises in the near future.
(Image source: blog.opsramp.com)
In fact, AIOps adoption is already gaining momentum. According to OpsRamp’s Top Trends in AIOps Adoption report last year, AIOps has become a critical ingredient for IT operations teams today.
68% of IT decision-makers are piloting AIOps to better manage the availability and performance of business-critical IT services. Of these, 73% are using AIOps capabilities to gain more meaningful insights from system-generated and monitoring-related alerts, and 68% to cut through the noise and determine the root cause of performance issues. The top benefits of AIOps cited by the 120 enterprise IT executives surveyed were the ability to automate routine functions and eliminate tedious manual tasks (74%), and the avoidance of costly service disruptions through faster recovery (67%). 58% also cited better anomaly detection, and 48% better causality determination.
(Image source: blog.opsramp.com)
Final Thoughts – Implement Artificial Intelligence for IT Operations in Phases
In the last two years, AIOps has grown from an emerging category to an IT necessity. Many successful companies are already using AIOps to automate and improve IT operations by applying machine learning to their data, and are pulling key insights from that data to drive strategic business decisions.
Gartner recommends implementing AIOps in phases. “Phased approaches of AIOps tend to the most successful,” former Gartner research director Vivek Bhalla explained at the Moogsoft AI Symposium in September last year. “What we’re tending to find is machine learning of event data and of structured data tends to be the lowest hanging fruit, and that’s no bad thing. If you’re familiar with that, use that as the entry point into embracing this technology.
Early adopters typically start by applying machine learning to monitoring, operations and infrastructure data, before progressing to use deep neural networks for service and help desk automation, said Bhalla. Gartner also suggests that enterprises should select AIOps platforms capable of supporting a broad range of historical and streaming data types to gain a composite visibility into IT entities – including applications, their relationships, interdependencies and past transformations – as this will provide insight into the present state of the overall IT landscape.
“It’s not just a case of thinking ‘we’ve installed it, job done.’ It’s a strategy, so you’re measuring this and evolving this,” said Bhalla. “You’re looking for new additional use cases, insights that you may not have conceived before. What I would say is, there’s no harm going for the lowest hanging fruit and then expanding from there.”