Phrases like “2.0”, “next generation”, etc. are somewhat cliché specially in our technology industry. We often see a trend nowadays where companies try to market certain enhancements on top of an existing technology as a game changer or something revolutionary and disruptive that would be a pathbreaking 2.0. Keeping that ethos in mind, today I would try to paint a clear picture on whether Kubernetes is actually enabling the next-gen multi-cloud, or is it yet another overhyped term?
Well, before discussing Multi-cloud 2.0, let’s quickly recall what is Multi-cloud 1.0. So, Multi-cloud 1.0 is the use of multiple public cloud service providers where the businesses mostly rely on Cloud management platforms (CMP) to manage different services, most of them being cloud-native, from a one-stop interface, often referred as “single pane of glass”. These CMPs somewhat simplify the management of multi-cloud deployments, without these platforms it’s almost impossible for businesses to manage multi-cloud in the long run.
Now in the 2.0 edition, where the idea is to pivot to a multi-cloud of a different kind altogether – a multi-cloud leveraging federated Kubernetes for management of containerized applications and data residing on numerous public clouds yet having awareness on each other. It might sound confusing, but, trust me, the federated Kubernetes architecture simplifies handling multiple clusters deployed on multiple clouds with capabilities of resource syncing across clusters. Automatic deployment sync ups across clusters running on many public clouds is possible in Kubernetes.
Another important benefit of multi-cloud 2.0 include increased availability through cluster replication mechanism in multiple public clouds at times of outage. In this architecture Kubernetes actually provides an abstraction layer above the Cloud providers, hence by definition it prevents the vendor lock-in issues and businesses will be able to replace a cloud provider if needed.
Adding to the magic, Kubernetes is open source (originally developed by Google and now maintained by Cloud Native Computing Foundation), but that doesn’t make it easy. Companies setting up Kubernetes be in cloud or on premise, would still need to do a pretty complex job. Initial setup of federated Kubernetes in itself is a tough task. Certainly, not to lose hope though! The best practices and tools for Kubernetes are emerging at a rapid pace with the evolution of talent in marketplace.
Does that mean Kubernetes is the panacea for all the limitations of multi cloud 1.0? Well, Container and Kubernetes technologies are still new and in early stages of adoption. I believe innovative products and services are key to achieving mass adoption of any new technology. Recent announcement from Google (and similar announcement from other Cloud providers), Anthos, a new cloud service with the ability to manage Kubernetes clusters across multiple cloud providers, is definitely encouraging.
In summary, Yes, Multi-cloud 2.0 is a reality and a better way to implement multi-cloud strategy is by using Kubernetes.