Multicloud computing has become the hottest theme in enterprise networking this year. As noted in this recent Mesosphere survey, multicloud adoption is doubling year over year and more enterprises are moving their workloads—many of them containerized microservices—to large-scale production multiclouds.
Multicloud isn’t necessarily tied to the concept of a “single pane of glass” to monitor and control it all. Instead, multicloud refers to the need for enterprises to avoid lock-in to a single cloud provider, to store and process data in two or more cloud platforms when necessary, and to flexibly shift workloads among them when doing so is more cost-effective and performs better.
As we inch closer to 2020, the centripetal forces that have been pulling workloads toward AWS, Microsoft Azure, Google Cloud Platform, and other public clouds have been waning as enterprises shy away from committing entirely to these monolithic services platforms. When enterprises consider the multicloud option, they may focus on using several specialized public clouds in lieu of or in conjunction with on-premises and private clouds to run various computing and storage workloads better, faster, and more cost-effectively than is possible with any one of the diversified public clouds.
This shift toward selective per-cloud workload deployment practice is encouraged by enterprise evolution toward microservices architectures, especially those involving Docker, Kubernetes, and other cloud-native platforms. Cloud-native microservices architectures facilitate the splitting and deployment of diverse workloads across specialized clouds. For workloads that aren’t performance-sensitive and don’t involve a continuous, high-volume flow of messages and data, it might make perfect sense for enterprise IT to run the associated microservices on disparate clouds, especially when each cloud has been optimized for specific compute and storage workloads.