This scenario is every engineer or IT leader’s nightmare for good reason: just a few seconds of downtime can negatively impact a business. Research shows 53% of shoppers expect e-commerce websites to load in 3 seconds or less, and 19% will leave if it takes longer.”

“The success of today’s digital businesses almost exclusively depends on the performance and availability of business applications,” says Mandy Long, former vice president of IT Automation at IBM. To meet customer expectations, you need to deliver continuous performance. But Long cautions that few businesses are positioned to do so consistently and cost-effectively.

IBM’s new series, “Rethink & Automate,” invites leaders to reimagine common business and IT processes by approaching them from a greenfield perspective and embracing automation. We asked Long how she would redesign cloud operations for a new company to ensure application performance.

Automation requires visibility, and a lot of organizations still have blind spots when it comes to seeing and monitoring their applications. Imperfect or limited visibility into fluctuating demand can result in under-provisioning or over-provisioning application resources. When you can’t predict and dynamically allocate exactly what resources your applications need, your engineers end up guessing.

End unmanaged and costly scale

To avoid being “that guy” who inadvertently took down an application on one of the busiest days of the year, cloud engineers might over-provision to mitigate performance risks. Multiply that across many applications and you get a lot of wasted spend. This common practice helps explain why organizations estimated that 32% of their cloud spend was wasted in 2022.² On the flip side, ballooning budgets and bills can trigger under-provisioning, affecting application performance and siphoning funds from other transformation efforts.

If your cloud spend is more than you’d like to admit, it’s because allocating just the right amount of resources to keep applications running well is incredibly difficult to do without automation technology that can monitor the dynamic needs of applications and keep them running with just enough.

You can’t make app-aware cost optimization decisions that account for resource needs and dependencies across the full stack—app to platform to hardware—without automation. The cumulative effect of automating micro-improvements to your allocations—at a rate that exceeds human scale—can result in major cost savings and continuous performance.

Opportunity for IT leaders

“Building on what Long said earlier, if I were to design cloud operations for a new company, I would invest in cloud cost optimization software that can automate resource allocation without compromising performance and operationalize it for continuous and lasting results,” says Program Director of IBM Automation, Asena Hertz. “There’s incredible upside for the business to get maximum business value for every dollar invested in public, private, hybrid and multicloud.”

As an example, Providence—a healthcare organization of 120,000 caregivers, 52 hospitals and 1,085 clinics—safely migrated to the cloud and achieved over USD 2 million in savings through optimization actions while assuring application performance—even during peak demand.³

Discover new ways to improve your application resource management with IBM Turbonomic

Another example comes from the IBM CIO Organization, which manages the applications that keep IBM running. Implementing 45,000 automated resourcing actions per month, it achieved a 3.8 TB decrease in cumulative memory limits and a 64% decrease in CPU requests.⁴ These efficiencies help reduce cost of labor as they eliminate time spent on manual remediation and free up time for innovation. They also help extend the lifespan of existing infrastructure and avoid unnecessary investment in new hardware.

Selecting the right tools

Cloud engineers and ITOps teams won’t execute cost optimization actions, let alone automate them, if there’s even a hint of performance risk. So, the software that continuously generates cost optimization actions needs to take a performance-first approach while accounting for the entire application stack and all the resource dependencies across the infrastructure it runs on.

Hertz recommends you look for these capabilities in any cloud cost optimization solution you consider:

• Efficiency without risk: Application-aware, full-stack analytics to ensure automation drives efficiency but never risks performance
• Trust: Automation that engineering, ops and app teams can trust–they always get the data they need to confidently execute actions, then adopt automation at scale
• Integrations: Automation that can be integrated with pipelines, processes and workflows
• Unified platform: One platform that supports cloud cost optimization across compute, volumes, database, Kubernetes and PaaS, including RIs and discounts
• Future-ready: Analytics that can be extended to future tech because it’s grounded in the fundamentals of IT, matching demand to supply across hybrid cloud, multicloud, edge and beyond