Cloud cost models based on 'inaccurate assumptions', claims research

Research points to assumptions that lead to surprises around cost savings

Customers across the channel are dipping their toes into cloud, with many migrating workloads to in the hopes of saving money. However, data released this week finds cloud cost models tend to be based on "inaccurate assumptions," leading customers to see less savings than expected.

According to Economics of Cloud Migration, a report by analytics platform provider TSO Logic, such assumptions include believing a cloud provider's hardware is similar to hardware being deployed on premises. The report points out that public cloud platforms may be newer with better price and/or performance.

Additionally, the report points to assumptions around hardware pricing leading to customers saving less than expected or, sometimes, finding out that moving to the cloud is actually more expensive.

"[Customers assume] that hardware pricing is basically the same for on-premise and cloud platforms when in reality public cloud providers benefit from massive economies of scale and often create custom hardware and software configurations," the report says.

The report also says customers tend to think their current on-premises resources are balanced and that their datacenters are using power efficiently, when this may not be the case.

According to the vendor, the biggest oversight, however, is customers' use of "incomplete ‘direct match' methodologies when projecting cloud costs."

"Baked into many cost models is the assumption that current on-premise resources are sized appropriately and that cloud instances should be provisioned exactly as they are provisioned on-premise," the report says. "New research from TSO Logic, however, reveals that most on-premise workloads - more than 80 percent - are currently overprovisioned."

The vendor encourages a "right-sized match" approach where provisioning is conducted using historical utilisation patterns and only for cloud resources that particular workloads require. It claims that through this approach, customers see an annual savings of at least 30 percent with the cloud.

"Organisations have tried to manually map their current environments to cloud, yet accelerated change and the sheer number of cloud compute options makes that impossible now," Aaron Rallo, CEO of TSO Logic, said in a statement. "Once you factor in modern compute capabilities, new service offerings and underused resources, organisations can slash their estimated cloud bill by… 36 percent, empowering them to focus on their core business with faster go-to-market execution and improved customer experience."