Continuous optimization (CO) is the process of using artificial intelligence and machine learning to drive the automation of infrastructure management in a way that responds in real time as workload needs change and cloud service offerings evolve.
Continuous optimization automatically responds as workload requirements and cloud service offerings change, runs the optimization recommendations made by infrastructure management solutions against appropriate internally- and externally-driven controls and policies, then operationalizes the changes required to keep infrastructure optimized.
The result of continuous optimization is infrastructure that is always tuned precisely to the demands of the workload and goals of the application.
Organizations implement CO with the intention of alleviating many traditional obstacles to efficient infrastructure management.
Cloud management is simply too complex for human beings to do well on their own. This complexity comes down to three areas of expertise that all IT professionals understand, but are impossible to master holistically at a deep and ongoing level.
First, comprehension of the generic requirements of workloads and applications is a deep area in and of itself. Compute, memory, storage, bandwidth and latency, placement, and database are areas in which one could individually spend an entire career. Even the most tenured IT professionals cannot have complete expertise across all layers of the infrastructure stack.
Second, although workloads can be similar, each is fundamentally unique and can evolve different requirements over time depending on a variety of factors and pressures. The engineering team who developed an application can usually predict and profile its requirements with the most certainly, but these individuals rarely also retain comprehensive and up-to-date knowledge of infrastructure options and quirks. And the realized demands on and performance of workloads in production change continuously—analysis of a workload’s requirements must be ongoing to be relevant.
Third, cloud service options expand and iterate as quickly as any workload’s requirements. With millions of possible total configuration options available across public cloud service providers at any time, no human can be omniscient.
CloudOps professionals are beholden to the demands application owners and users. Continuous optimization enables IT Operations to manage infrastructure in ways that is completely respectful of and driven by internal and industry policies, leveraging appropriate approval workflows (as driven IT service management—ITSM), change windows, and other best-practice constraints.
Just like manually-driven cloud workload optimization, continuous optimization better matches workloads to infrastructure, ensuring peak application performance, and reduces cloud costs by maximizing each cloud instance and avoiding unnecessary expenditure. The differentiation is that continuous optimization responds automatically and can dynamically make permitted changes without human intervention (if allowed).
If your organization already practices Continuous Integration and Continuous Delivery, the natural next step is Continuous Optimization: CI/CD/CO—a third step in your pipeline that ensures your cloud apps are always optimized.
Optimization as code is the use of infrastructure as code (IAC) tools like Terraform and CloudFormation to empower workloads to continuously optimize themselves by ceding infrastructure selection responsibilities to machine learning and artificial intelligence.