Is cloud cost out of control?

Recent reports indicate cloud cost control is becoming a major issue for businesses. Focussed on moving storage and compute quickly to the Cloud, drawn in by attractive pricing created by competition between providers, many businesses are now surprised at the spiralling cost.

Despite a determination to cut infrastructure expenses, spending on public cloud is now a significant slice of corporate IT budgets and it is the on-demand nature of its supply that is causing the problems. Three particular cloud characteristics are chiefly responsible for the waste:

1 - Cloud pricing complexity

What began as a simple pricing model has developed a plethora of options, with few understanding the cost implications of all the choices made. Instances can vary in price depending where they run, or whether they are the newest version and choosing the wrong storage class can add real cost too.

2 – Instance size matters

Provisioning instances and sizes correctly requires experience of cloud environments. Migrating from on-premise, an organisation must first understand how they are truly utilising each workload to accurately evaluate cloud options, or typically they will overprovision and that’s hard to overcome.

3 – Cost implication transparency

When provisioning, resource owners rarely understand the full cost implication of running their applications in the cloud. On first consideration the hourly cost for instances may look inexpensive, but when overprovisioned and running for a year (8760 hours), it’s easy to lose control of the costs.

Optimising and managing your cloud costs

When utilising public cloud, it is essential businesses reduce waste and the associated cost, which globally has been calculated to run into billions of dollars. They need to adopt cloud cost optimisation best practices to reap the full benefits and here we highlight a few of the best.

Identify unused or unattached resources

Finding and removing unused or unattached resources you are being charged for, is the easiest win in the optimisation game. Typical examples include forgetting to turn off the temporary server used for a particular function, or when terminating instances, forgetting to remove the attached storage.

Combine resources

Combining computing jobs into fewer instances will avoid being charged full price for an idle instance with a low CPU utilisation level. Low utilisation is now a thing of the past thanks to the ability of the cloud to allow you to scale up your computing power at any time to meet an increase in demand.

Right size computing

Application owners will typically over-provision as they try to avoid resource shortages that can impact performance. Moving on-premise hosted apps to the cloud, owners will often just replicate their existing VM resource allocations, without first calculating the resources the app actually needs.

Reserved instances

If your future is committed to the cloud, then you should seriously consider reserved instances (RI). Signing up for one or three years you can save up to 72% with RI against just pay-as-you-go pricing, but combining the two will help manage costs across predictable and variable workloads.

Azure low priority virtual machines

Azure low priority VM are compute instances allocated from a date centre’s spare capacity and they are offered at a highly discounted rate compared to on demand services. Bidding for these for the right workloads can deliver serious cost savings, but you will lose them if someone bids higher.

Do not hoard data

Keeping every file your business has ever created is no problem on your local machines, but in the cloud it all adds up to real costs; real ongoing monthly cost whether you access that data or not. Consider storage and choose the most appropriate type that balances functionality and cost.

Multi or single cloud

Multi-cloud options help avoid vendor lock-in, a laudable outcome, but whilst you might improve availability and uptime, you risk missing out on the potential volume discounts offered by a single cloud vendor. And you must also consider the cost of traffic between clouds and additional training.

Doing nothing is not an option

Hopefully, the best practices outlined above will help you optimise and manage the ongoing cost of your cloud services. But it’s important to remember that the cost-savings you were promised by moving storage and compute to the cloud is possible if you manage it carefully.

If you take the promise at face value and ignore the underlying need to look closely at your current usage and requirements, the potential spiralling costs of cloud will not only surprise you, but risk you having to curtail your future activities.

Which optimisation strategy you choose is largely unimportant, as long as you do something. And if you are not sure what approach will best get your costs under control, we are only a call or an email away, ready to offer you the benefit of our years of experience and undoubted cloud expertise.

Add your comment

The content of this field is kept private and will not be shown publicly.