Enterprises across the globe made a hasty leap into the cloud when the Covid-19 pandemic hit, rushing in a panic to get products and services online as fast as possible as the world shut down. But now, with big bills rolling in and calmer heads prevailing, companies are looking to pivot from a cloud-first to a cloud-right approach.
Rather than simplifying operations for enterprises, the quick move to the cloud during the pandemic “created chaos” in both their operating environments and operating models, Aparna Sharma, GM of Hybrid Cloud Services at IBM Consulting, told Silverlinings.
Now, enterprises have found themselves saddled with “very complex environments” and a lack of interoperability, with multiple workloads scattered across multiple clouds.
Indeed, back when the pandemic was ramping up in June 2020, IBM predicted organizations would be using at least 10 clouds by the end of 2023. It reiterated that forecast in a hybrid cloud report published in mid-2022.
“With that chaos and architectural complexity and the cost of the public cloud, I think organizations are now seriously stepping back and thinking about what’s the right way to proceed forward,” she explained.
Some of this thought process is reflected in the discussion around repatriation of cloud workloads to an on-premises environment. But Sharma noted that among the enterprises IBM Consulting works with, 70% of mission critical workloads have still not yet moved to the cloud. That’s due to a range of factors such as compliance and data residency requirements, costs and innovation in the private cloud sector.
Now, as opposed to three years ago, Sharma said there’s a lot more thought going into enterprise evaluations of whether and how to move more of their workloads to the cloud, which cloud to use and whether to bring other applications back on-prem.
Among other things, Sharma said enterprises are beefing up planning efforts, looking at things like what data governance policies need to be in place, what tooling is needed for different kinds of migrations and how best to test their applications post-migration to ensure data completeness and quality. They’re also pushing to build platform agnostic capabilities to avoid vendor lock in.
Additionally, enterprises are working with the likes of IBM to build their own analytics to better measure what they’re spending on the cloud. While it might seem like a straightforward question to answer, Sharma noted many enterprises lack a holistic view of the cloud services different departments are using, leaving them vulnerable to surprise bills.
“We are building out analytics around the FinOps-based cloud economics for different technologies, workloads and building out essentially ROI models for them,” she said. “Instead of this being just an emotional decision for them, making a much more quantified decision from a cost standpoint.”
In addition to costs, Sharma pointed to architectural challenges, people and operating models, security and partner ecosystem challenges as the primary hurdles enterprises face in terms of the cloud.
“Not only do enterprises have these cloud services partners, but they also have let’s say a GSI [global systems integrator],” she explained. So, they’re trying to figure out “how do you develop a unified vision across all of these partners so they’re putting the good of the organization across the siloed pieces.”
And last but not least, Sharma said organizations are weighing what to keep on their own mainframe, what to push to the cloud and how to make both of those environments play nicely.
“These organizations are still relying on the mainframe for critical applications, you can think about anything from credit card processing to core banking to reservations, whether it’s financial services or healthcare…because there’s no cloud environment that’s able to meet the eight 9s” of reliability, she concluded.
Have a story to share with us? Write to the editors here.