Why autonomic intelligence is the AI the cloud needs

While we hear a lot of discussion about “hybrid cloud”, the truth is that the majority of organisations aren’t there yet.

A truly hybrid cloud would allow workloads to switch seamlessly between cloud environments, whether public or private, depending on the needs of the organisation.

Instead, what the vast majority of organisations have now is “multi-cloud”; while they have a number of environments, both private and public, and workloads may fluctuate in size and demand, barring exceptional circumstances those workloads will stay where they were first placed.

However, this doesn’t mean that organisations are taking the easy route. Issues with multi-cloud alone suggest that a truly hybrid approach is a long way away.

A very unique skillset

For instance, in a recent survey the majority of organisations admitted that they don’t have the skills needed to manage a multi-cloud environment.

This is quite believable when one considers that by their very nature multi-cloud environments aren’t managed from a single user interface.

An organisation’s private and public clouds have, in most cases, come from separate vendors, each working differently, with their own management tools and interfaces.

>See also: The pros and cons of hybrid cloud migration explained

This adds extra complexity on top of an already potent mix of challenges, such as deciding which workloads should be placed in which cloud; balancing cost and performance across all environments; meeting service level agreement and quality of service (QoS) obligations; and actually choosing between the universe of the different cloud vendor options available to make sure you have the best for your business.

Even if the IT team is one of the minority with the right skills to solve all of these challenges, and can implement a multi-cloud that meets every single business need, it will still find itself expending a huge amount of time and effort to get there; in turn taking resources away from other areas.

This is the best-case scenario; at worst, organisations will find that their multi-cloud doesn’t meet the needs of the business. For instance, workloads may not be correctly balanced, costing the business either in productivity or increased financial costs.

Alternatively, IT may not be able to meet the budgetary or QoS needs of the business that are often the main reason to pursue multi-cloud. If IT teams can’t meet these challenges themselves, and can’t be sure that their cloud vendors will solve them, what can they do?

Operating on instinct

Multi-cloud management needs to be instinctive; the needs of the business and industry pressures change constantly, and attempting to consciously identify and react to them would lead to IT teams constantly running to stand still.

Instead, the multi-cloud environment should be an autonomic system, able to adapt in real time to stimulus and environmental pressures to ensure it always has the best performance available.

>See also: How to do hybrid cloud integration securely

Much as our bodies adapt without us making any conscious decision when excited, or frightened, or even just hot in order to help us survive, so should the cloud.

However, we need to know that our instincts are correct. For instance, if your body immediately began to sweat more as it got colder, your chances of survival would be limited.

Similarly, a multi-cloud environment that constantly, and automatically, increased the resources available to the organisation could create a financial and management black hole if left unnoticed.

Ultimately, it would repeat the old problems of virtual and cloud sprawl – just in a more efficient manner.

The right system

Autonomic decisions need to be intelligent and driven by an underlying understanding of the factors impacting the business.

Quite simply, if IT teams have to continually monitor the multi-cloud environment to spot errors then much of the value of such a system is gone.

The best underlying understanding is an economic one; if the system follows a market-based model, with strict budgetary constraints and rules, then its nature will always be to find the proper equilibrium between application demand and cloud supply.

In such a system, the multi-cloud environment would be treated as an economy in its own right; with resources bought and traded depending on their applications’ – or end-users’ – real-time needs, while obeying the organisation’s own budgetary rules.

>See also: The 5 phases of overcoming hybrid cloud data integration

The multi-cloud could then become truly autonomic, and IT teams could be freed up to focus on providing value to the organisation, rather than having to pay conscious attention continuously to managing multiple cloud environments.

In an ideal future, organisations would use a system built on autonomic intelligence, leveraging market-based algorithms to manage multi-cloud environments and ultimately deliver hybrid cloud, with workloads always placed properly to suit business needs and provide peak performance.

However, in advance of hybrid cloud – even in a multi-cloud or a single private cloud architecture – such a system is as vital as the human body’s own autonomic system. Quite simply, unless they can trust their system, IT teams will never be able to trust their ability to meet the challenges of the cloud.

 

Sourced by Charles Crouchman, CTO at Turbonomic

Avatar photo

Nick Ismail

Nick Ismail is a former editor for Information Age (from 2018 to 2022) before moving on to become Global Head of Brand Journalism at HCLTech. He has a particular interest in smart technologies, AI and...

Related Topics

Hybrid Cloud