According to Research and Markets analysis, the multi-cloud market will grow to USD $44 billion by 2022. 451 Research says 69% of enterprises will have multi-cloud and hybrid cloud IT environments by 2019. Containers are one of the key enablers for cloud adoption, with Kubernetes as the orchestration solution of choice propelled by a large open source community delivering new innovations at a rapid pace.
A couple of months ago, as AWS joined Azure and GCP in announcing their managed Kubernetes container management solution, at Nirmata we announced our support for EKS, and also recently we demonstrated seamless management of GKE clusters with Nirmata.
In our conversations with customers, it is clear that enterprises want to use different clouds for best-in-class offerings for specific workloads. And private cloud infrastructure also continues to grow, as there are applications that will be run on dedicated / bare-metal infrastructure for security, cost and performance reasons.
As enterprises move towards deploying containerized applications and get comfortable with the portability containers offer, they need a unified enterprise-grade management plane across public and private clouds. Here are some key considerations that we have identified for such a management plane –
- Consolidated Operations – As containers have become popular with enterprise developers, there have been many instances of siloed projects driven by business units within an enterprise. Now, IT operations teams are being asked to bring these back under a single umbrella to apply common policies, governance, and manage costs. There is an additional element of complexity with containers as packaged applications are blurring the lines between infrastructure and application operations teams. What is required is a single platform to provide the necessary infrastructure and application management for the stack.
- Built for multi-cloud – As container adoption progresses, there are certain workloads more suited for public clouds while others need to stay on-premises. Cloud providers also offer their own managed container services. While having choices is great, enterprises need to think about how to manage all of these under single set of workload policies so that developers don’t have to worry about underlying infrastructure.
- Governance – Enterprises have many requirements around industry compliance, regulatory requirements and their own policies under which the technology infrastructure needs to managed. And this governance needs to be a common and ubiquitous layer across clusters and across clouds.
- Isolation – A key value proposition of Kubernetes is the ability to share infrastructure across heterogeneous workloads. However, this requires proper management controls from the operations perspective to meet required SLAs for different environments across performance, security, compliance and risk management considerations. Nirmata delivers this isolation for a multi-cloud infrastructure through constructs of environments and multi-tenancy. More on this in a later post.
- Workload Portability – Kubernetes and containers technology make application portability easy, but when you are deploying applications across on-premises infrastructure and different clouds, you need to take into consideration the differences in underlying infrastructure. A unified cloud-native management plane that understands the underlying infrastructure is best positioned to deliver on the promise of application portability. On a related note, Nirmata enables application portability today, we will be shortly announcing partner integrated solutions to manage both application resources and the underlying data across clouds, hence delivering a comprehensive application portability solution.
- Full Stack Visibility – Enterprises need a single view across clusters and clouds to drive necessary integrations for CI/CD, security, logging and monitoring. A unified approach to application lifecycle management and performance monitoring is essential for success.
Learn how Nirmata helps deliver unified management plane for your containerized workloads.
What are your key considerations for multi-cloud workload management for Kubernetes?
As always, your comments and feedback are highly appreciated.