The AIOps promise: future of IT management across the hybrid data centre

The rise of AI and machine learning has opened up a whole new world of possibilities for dealing with this level of complexity. The global data centre automation software market is predicted to reach a value of $9.42 billion between 2019 and 2023, with a CAGR of 22%*. Data centres of the future will continue to augment the management by humans; the industry is on a path towards more automation of these environments. And the journey between the status quo and that automated future is being fast tracked by the application of artificial intelligence for IT operations: “AIOps”, of which there has lately been much hype.

 

What is AIOps?

 

Gartner’s definition** for AIOps is:

 

“Platforms that utilise big data, modern machine learning and other advanced analytics technologies to directly and indirectly enhance IT operations (monitoring, automation and service desk) functions with proactive, personal and dynamic insight. AIOps platforms enable the concurrent use of multiple data sources, data collection methods, analytical (real-time and deep) technologies, and presentation technologies.”

 

The drivers for AIOps

 

The future of IT operations is autonomous: enterprises need IT operations that never fail to deliver mission-critical services, that adapt to business innovation, and consume resources ever more efficiently.

 

Achieving this is difficult because the enterprise data centre today is made up of multiple infrastructure silos with disparate element managers that cannot adequately correlate or communicate with each other. The amount of complexity and moving parts in today’s data centre is staggering and IT operations and infrastructure teams simply do not have the visibility, capability or tooling to orchestrate across silos and private/public clouds.

 

As newer technology is deployed, the need for visibility across the entire environment becomes ever greater. The ability to have cross silo transparency and automation coupled with unifying and streamlining this mass of data, starts to tackle the numerous challenges that have arisen as a result of this limited silo structure.

 

Why AIOPs is different

 

If we compare running a data centre to driving a car, most monitoring tools provide a post-process analysis of why a problem occurred across a certain layer of infrastructure, or to use the car analogy, why your car crashed! The insight that AIOps provides ensures alerts are delivered in advance of the impending accident so that incidents can be anticipated, and through machine learning and algorithmic learning, intelligent action can be taken and the course corrected ahead of an accident ever taking place. The massive expense, disruption to customers and damage to an enterprise’s reputation should an outage occur, make AIOps capabilities in today’s hybrid data centre essential.

 

For a data centre to be truly autonomous, a massive amount of information must be collected about its operations and infrastructure. This is where the years of machine learning underpinning AIOps comes to the fore. It enables AIOps to discern and prioritise from thousands of alerts, essentially cutting out the noise. AIOps intelligence as a result of machine learning is also relevant in the context of providing a performance baseline, as well as delivering vital analytics for both applications and delivering business SLAs.

 

AIOps benefits and use cases

 

AIOps significantly reduces or removes mean time to resolution, preventing outages and enabling massive cost savings for the business due to increased uptime.

 

For operations teams, AIOps provides visibility and insight across silos and allows IT to become more business focussed, allowing them to innovate rather than simply operate, as has previously been the case.

 

True AIOps can enable an end-to-end real-time visibility, allowing for utilisation and capacity planning, which entirely mitigates infrastructure overprovisioning, delivering massive CAPEX and OPEX savings.

 

Moreover, the detailed insights enabled by AIOps can help increase the profiles and credibility of IT teams and application owners, shifting from a situation of finger pointing and “war room” scenarios, to providing valuable intelligence and reporting on the capacity, health and utilisation of the hybrid infrastructure.

 

AIOps’ ability to simplify data centre operations is vital; with cloud migration, there is a perceived risk in placing infrastructure in the hands of public cloud vendors. With AIOps, cloud vendor choice becomes less of an issue, as the entire heterogeneous infrastructure is controlled by the owners of IT, rather than a faceless public cloud vendor.

 

Barriers to AIOps adoption

 

AIOps enables strategic monitoring and control, ensuring application and infrastructure SLAs are met. One of the major barriers to its adoption is lack of trust and resistance to change. Organisations do not like moving away from the technology and tools they are familiar with. But simply put, without automation, managing the complexity and multiple moving parts of the hybrid data centre is becoming overwhelming for operations and infrastructure teams.

 

One of the core propositions of AIOps is “handing the keys” (and the control) of the infrastructure back to operations teams. Much as with a commercial aircraft, the flight controls can be set to automatic, but the pilot is free to take over at any time. It is ultimately up to the operation teams as to how much they wish to automate.

 

AIOps is more than an insurance policy against risk, it is a pivotal enabler of digital transformation, allowing companies to powerfully shift from reactive trouble shooting to proactive optimisation, streamlining operations to better serve their business and deliver organisational agility.

 

*Global Data Centre Automation Software Market – Technavio, June 2019

**Gartner definition of AIOps