On-demand cloud services provide the agility for enterprises to operate at scale, although that cost may run up when deployment grows. This calls for companies to optimize cloud spending, which can be achieved through a few simple adjustments without impacting performance, write 451 Research’s Owen Rogers, Research Director, and Jean Atelsek, Research Analyst. Here are their top tips for how to optimize your organization’s cloud services.
Find the best deals on Cloud Services
There is the typically overlooked fact that, when businesses beef up their cloud capabilities, the on-demand rates for cloud services are usually the most expensive. This can be addressed by purchasing capacity in advance, or enterprises could set a level of monthly spending for predictable workloads with longer-term baseline capacity.
Findings from 451’s Cloud Price Index across the five major cloud service providers shows that average cost savings of 36% can be made from just discounts afforded by a longer commitment – this translates to about $24bn of savings on direct cloud costs annually!
Moving toward a longer commitment does not spell an end to flexibility. Businesses can always supplement with on-demand resources as needed to cope with fluctuating computing demands.
Key providers such as AWS, Microsoft Azure and Google Cloud Platform offer their own savings plans for aggregated purchases with a range of flexibility. For example, the AWS Savings Plan requires a commitment to an hourly level of spending over a one- or three-year time horizon for a plan that allows switching of instance types according to usage levels.
Such bulk deals help hyperscalers to generate better cash flow. Enterprises benefit from better budget planning and savings, although they have to be mindful, since rollover provisions will not be available if the cloud resources are unused.
The Cloud Complexity Storm
Get dynamic scaling
Another approach would be for companies to scale resources dynamically through rightsizing. Cloud is, after all, inherently flexible, and autoscaling is a built-in feature for most cloud services.
Enterprises can set up autoscaling during the initial deployment, so that the VM will grow and shrink compute capacity following a schedule or in response to spikes and troughs in demand.
Additionally, enterprises can opt for tools from third-party vendors to gain better visibility and control of their cloud spending. Some of the options available in the market include:
- Turbonomic (acquired by IBM)
- Cloudability (acquired by Apptio)
- CloudHealth (acquired by VMware)
Mix and match cloud providers
Meanwhile, some enterprises are pursuing a multicloud model for cost arbitrage. This can be challenging. Customers have to be vigilant over differences in compute pricing and dynamically tune an application to an availability region with the lowest pricing.
The three major hyperscalers – AWS, GCP and Azure – offer low-cost ephemeral instances at steep discounts (80% or more) compared with on-demand rates. The caveat remains that the instance can be terminated without warning when the hyperscaler needs to reclaim the capacity.
Cloud-native software vendors, service providers and open source projects can provide some help for enterprises managing a distributed cloud platform. CAST AI, for instance, uses a single stretched Kubernetes cluster to move workloads between and across different clouds based on minute-by-minute or penny-by-penny changes in price and availability.
Optimize instead of switching cloud providers
Enterprises that want to cut costs must keep in mind that optimizing is a far better option than switching providers. Switching providers can lead to the hassle of managing differences in product sets and capabilities, which is not worth the savings. If it isn’t broken, don’t fix it.
Want insights on cloud computing trends delivered to your inbox? Join the 451 Alliance.