Why the future of cloud is autonomic
Thu 8 Sep 2016
Charles Crouchman, CTO, Turbonomic, argues why companies need to embrace an autonomic approach to IT management…
Cloud is still in really early stages of development and application in enterprise. Looking beyond shadow IT, which was about being able to spin up a VM in minutes rather than it taking weeks or months for internal IT to provision, large organisations are just starting to take steps into the cloud. Typically they have started in test and development infrastructure environments – and in many cases that’s still where cloud is today.
It is relatively uncommon to see a fully rolled-out production cloud deployment, except in cloud-native companies which have built their businesses in the cloud from the get-go. Most traditional organisations, however, haven’t been in this position – meaning that across enterprise the cloud remains in its infancy.
For companies embracing the move to cloud, a multi-cloud approach is typical before advancing to a hybrid strategy. Multi-cloud is managing workloads in multiple clouds discretely, whereas hybrid is the vision of bursting workloads from private to public and back as demand peaks.
The cloud-native compromise
Cloud is likely to stay in the multi-cloud stage for a while before hybrid starts to take over. It isn’t easy to take traditional applications and make them cloud-native, or even capable of taking advantage of what the cloud offers.
At the same time, smart organisations will realise that not every app is suited for every cloud – there are differences between apps and clouds, and it makes sense not to be locked solely into any one cloud.
In this way multi-cloud has and will continue to have ownership over early adopters; but as organisations experience more from the cloud, they will soon come to appreciate the benefits of control in their own private infrastructure, leading to a growing interest in the role of hybrid.
The cheapest cloud isn’t, by definition, the best option
The differences and vast expanse of choices in service is almost overwhelming. What’s more, the only variable that is fully transparent in the cloud is price. Ask any IT professional who has tried to speak with Amazon about an SLA and you’ll hear how assuring performance is a major challenge. For some of the leading cloud providers, like AWS or Google, focus is far more on the consumer side than on understanding the enterprise. Recently Google was smart enough to hire Diane Greene, who founded and built VMware. Greene heads up Google’s cloud offering and her experience means that she knows enterprise extremely well; but the tech giant’s core DNA remains consumer-focused.
The rationale of price-cutting in Cloud
To add to the puzzle, there is always so much talk about specific price cuts, but less about the ways costs rise. No business can cut costs forever, and often they cut costs just long enough to get users locked in so that they have a hard time leaving when prices inevitably rise.
The other point is that you get what you pay for, and enterprises who focus solely on cost or price will be sorry when an outage loses them far more in business than they ever saved through their package.
The cloud needs power, cooling, racks – they just belong to someone else, operating at a scale that provides efficiencies which other companies can’t match. While this may be true in some cases, if you consider all the important pros and cons companies must consider – particularly performance – the cheapest cloud isn’t, by definition, the best option.
The self-regulating cloud
These trade-offs in the cloud between price, cost, choice, performance, resiliency, and many others are difficult to manage on their own, let alone having to manage the impact each has on the other. It is not surprising that organisations are keen to explore autonomic systems to help manage these complex environments.
IT generally requires highly-trained, very talented professionals to monitor the many thousands of activities taking place in a data centre or across cloud infrastructure. Staff have to think about provisioning VMs, how to size them, where they should be placed and how they should be configured. In a cloud of any size it is problematic to manage these requirements with a human-based approach.
Just as the human body needs an autonomic system to control heart rate and other vital biological functions, such as breathing, data centres and cloud configurations can also be managed autonomically – with no thinking necessary. This freed-up talent allows companies to unleash innovation and focus on building IT for their business, rather than worrying about whether to turn the lights on or off.
