CIOReview | | 19 DECEMBER 2022consumption, and therefore cost, to provide a more holistic view of spend across our multi-cloud estate. This approach leverages observations from the Theory of Constraints on favouring system optimization over local optimization as we have many talented engineers optimizing the costs of their application account, but that capability is not yet ubiquitous in the organization. FinOps tooling, and the associated process, provides cloud centers of excellence with the ability to be more consultative with application teams, guiding them on how to balance capacity, performance, and availability with cost. It also gives the ability to steer workloads, reservations, and commitments at the cloud provider level, optimizing for the greater good.Multi-cloud has compounded the existing complexity of many enterprises operating environments, with customer journeys spread across multiple providers. Growing customer expectations regarding response times and the move to mobile as the device of choice for many means that any incidents need to be diagnosed and remediated as quickly as possible. This requires a comprehensive approach to logging, monitoring, and alerting that joins up information from all providers, including the private cloud and legacy environments, to aid recovery and communications to customers. This requires machine learning or AI to make sense of the firehose of data coming out of each provider and surface relevant information to engineers or automated systems for action. An operations data lake and AI are table stakes for hybrid-cloud observability.Seamless connectivity amongst all providers, public and private, is required for hybrid-cloud, allowing developers to place their workloads where business requirements are best met within cost and risk appetite via a toolchain. Networking complexity is something often glossed over by those pushing the concept of a provisioning single pane of glass and is often where those solutions are undone during execution. The introduction of a carrier neutral facility capability or a software defined WAN are good first steps toward a unified network fabric for the hybrid cloud. These solutions allow the externalization of certain key capabilities, like API gateways, global traffic managers, and data lakes or processing, that increase portability and resilience across multiple regions or cloud service providers. Integrated CI/CD toolchains require this sort of connectivity to empower developers with the capability to deploy their code to the appropriate provider without having to take on the cognitive load to account for firewalling, LANs, WANs, and load balancers.These three capabilities are a good start and provide a way for enterprises to begin the move from multi-cloud to hybrid-cloud. In the meantime, the market will continue to mature things like container as a service platforms, networking, and security, and application teams can upskill their domain driven design, test driven development, and CI/CD capabilities. The full value of the hybrid-cloud approach will not be realized without application teams writing cloud native software that takes advantage of containers and/or functions as a service and the new resilience models they offer. Smart, robust deployments of FinOps, observability, and seamless networking with updates to processes and supporting operations will enable enterprises to realize value, continue automating away toil, and improve resilience as they continue their cloud journey. A single cloud provider with multiple availability zones in multiple regions meets most requirements for resilience and scale
< Page 9 | Page 11 >