Cloud customers are wasting money by overprovisioning resources
And all that energy feeding those humming servers isn't free either
Organizations are overspending on containerized workloads in the cloud by overprovisioning the resources needed, and could potentially cut costs by as much as 60 percent.
That's according to a report from cloud monitoring and optimization biz CAST AI claiming organizations on average provision a third more cloud resources than they end up using, which is not only a waste of money for the customer involved, but may also result in waste of the energy consumed by underused resources, CAST AI claimed.
This could be a growing problem in future, as enterprise spending on cloud-based infrastructure is expected to overtake the cash spent on traditional IT hardware, although growth was recently found to be slowing following a rapid rise during the pandemic.
CAST AI develops a platform to monitor customer use of Kubernetes cluster resources across the three major cloud platforms – AWS, Microsoft Azure and Google Cloud – and compares the resources an application is using with those that CAST AI calculates that the workload actually requires.
Its findings in this latest report are based on an analysis of data from thousands of clusters running cloud-based applications.
CAST AI found that on average 37 percent of the CPU resources provisioned for cloud-native applications are never used, and claimed that by optimizing clusters and removing unnecessary compute resources, its platform would allow those resources to be freed up and used for other purposes.
By applying a price arbitrage policy – swapping high-cost virtual instances with equivalent lower-cost machines – the impact of rightsizing is almost half of the cloud compute bill, according to the company, and would save customers about 46 percent of their costs in dollar prices.
CAST AI also analyzed the impact of moving workloads to spot instances where applicable, which - as Reg readers know - are virtual instances that take advantage of excess capacity that is currently unused by the cloud provider and available at less cost than standard On-Demand instance pricing. The potential savings here amount to about 60 percent compared with taking no action, it claimed.
AWS, Google or Microsoft – doesn't matter
According to the report, these findings hold across all three of the major cloud platforms, with a variation of plus or minus 5 percent. CAST AI said that it also does not depend on the application size, with a variation of less than 5 percent between small applications using $1,000 of resources per month and larger deployments of $100,000 per month. This would indicate that the rightsizing problem is universal and relates to how cloud-native applications are managed, the company said.
CAST AI said over half of organizations blamed a lack of visibility into cloud usage as the main reason for this wasteful behavior with regard to cloud resources, citing figures from an Anodot 2022 State of Cloud Cost survey. Addressing this typically requires a third-party solution since the monitoring tools from cloud providers do not display cost data in real time or have insufficient options for data sorting, the company claimed.
- US commerce bosses view EU rules as threat to its clouds
- Fortinet's cloud firewall ditches custom ASICs for Amazon's Graviton CPUs
- Broadcom CEO says hiking VMware prices is not his strategy
- Microsoft extends Azure Hybrid benefit to some on-prem software
The CAST AI platform provides free analysis for organizations to determine how their cloud resources are provisioned, but only paying customers have the option to let the platform take action based on its findings and right-size their cloud resource provisioning to match workloads to the optimum resources.
A previous report from the company highlighted that cloud overspending often occurs due to an over-cautious approach from customers when faced with a lack of information regarding how resources are being used.
At the time, independent analyst Clive Longbottom told The Register that using tools such as CAST AI made sense so long as the potential savings outweighed the cost of the subscription. However, he cautioned that customers would be relying on CAST AI to optimize their cloud costs, and should have a backup plan in case it should become unavailable in future. ®