How to Shift Workloads Across Clouds—and Simplify Your Infrastructure in the Process
Jul 16, 2019
Kubernetes has garnered the confidence of developers worldwide and now it’s in the process of revolutionizing how companies apply IT infrastructure to the compute problems they need to solve. Kubernetes transforms traditional IT operations into “infrastructure as code,” making deployments repeatable, agile and secure. Overall, Kubernetes makes distributed computing within clusters easier.
However, multicloud and multisite deployments introduce significant additional complexity for organizations operating Kubernetes at -- or beyond -- enterprise scale. Business needs are pushing IT teams to move workloads between sites and clouds, and to rapidly be able to clone, burst and then destroy workloads anywhere around the world, as needed, where needed and when needed.
But Kubernetes deployment across multiple regions, sites and clouds remains challenging. There have been several attempts, but nothing has yet solved the problem of running a distributed system of distributed systems, a cluster that spans multiple clusters.
Juniper’s multicloud storage and compute solution, Juke, removes the infrastructure silos between on-premises, campus, branch, AWS and GCP cloud resources with a policy-driven management platform for containers and persistent storage that automatically manages tiering, locality, availability and resiliency. Juke effectively adds situational awareness to a multicloud infrastructure.
Juke’s resilient, globally distributed storage layer allows organizations to operate workloads on-premises, across multiple public clouds and all the way out to the edge using a single, unified storage layer that transcends infrastructure boundaries. With Juke 2.2, Juniper Networks has delivered a solution to the problem of operating multiple Kubernetes across hybrid multicloud environments by providing persistent storage for container workloads wherever they need to operate. The key feature that Juke brings to the party is a situational awareness of latency across the stretched infrastructure. Juke is able to automatically adapt to the deployment situation to deliver the best app performance across a resource mesh.
Imagine a truly multicloud application capable of growing and shrinking on demand and even of recognizing when there is a problem with a given cloud, or region, and shifting workloads to other clouds to compensate. Now imagine that you don't have to rewrite your application in order to achieve this. Juke 2.2 and Kubernetes is a game changer: the multicloud uber cluster.