Introducing Hyper-Converged Kubernetes

By September 10, 2018 June 26th, 2019 Blog, Databases, Databases, Kubernetes

Kubernetes is the de-facto standard for container orchestration for microservices and stateless applications. Building upon more than 15 years of experience of running production workloads at Google, presumably the most challenging production environment in the world when it comes to scaling and meeting SLAs, Kubernetes brings incredible agility to your private clouds. Kubernetes provides several building blocks needed to build a production-ready private cloud. By design, it is not a turnkey solution because it is designed to provide the flexibility to its users to select their preferred components. For example, selecting which storage to use or which networking solution fits their needs the best.

On the other hand, Hyper-converged Infrastructure (HCI) provides a turnkey solution for your applications by combining virtualized compute(hypervisor), storage, and network in a single system. Consequently, HCI eliminates the complexity of integrating infrastructure components by providing an out-of-the-box solution that runs enterprise applications. The key design principle here is the turnkey simplicity.

Hyper-converged Infrastructure

Wouldn’t it make sense to combine Kubernetes and the principles of HCI to bring this turnkey simplicity to Kubernetes? What if, similar to HCI, your Kubernetes came with all necessary infrastructure components out-of-the-box? And imagine a native integration, which not only simplifies Kubernetes setup and management but also unlocks new use cases and workloads to run seamlessly.

These questions were the driving force behind our architectural decisions. Through hundreds of customer interactions focused on understanding their pain points, specifically around managing data-heavy applications such as big data, databases, and AI/ML, it was very clear to us that Kubernetes has a massive role to play. However, as mentioned before, Kubernetes is not the turnkey solution most of these customers were hoping for. Above all, they simply wanted a solution that works out-of-the-box to solve the deployment and lifecycle management complexity for their data-heavy applications. They didn’t necessarily want a “Kubernetes solution”, but a solution that specifically addresses their needs. Kubernetes just happens to be a key ingredient.

Introducing Hyper-converged Kubernetes

Hyper-converged Kubernetes is a software-defined application orchestration framework that combines containerized storage, networking, compute (Kubernetes), and the application management layer into a single system. As the name suggests, we bring together the simplicity of the hyper-converged infrastructure and the agility of Kubernetes.

Hyper-converged Kubernetes Concept

The idea is to use the building blocks provided by Kubernetes to achieve a turnkey solution similar to hyper-converged infrastructure. Most importantly, all infrastructure components are natively integrated into a cohesive system with Kubernetes at its core. The end result is a hyper-converged solution that is simple to use, while also providing the agility and scale of the container technology.

ROBIN: Industry’s First Hyper-converged Kubernetes Platform

ROBIN is the industry’s first implementation of hyper-converged Kubernetes. It brings self-service App-store experience and increased efficiency to Big Data, Databases and AI/ML. The hyper-converged Kubernetes architecture provides built-in storage, compute, and network, and an application management layer to control them. This enables DevOps and IT operations to leverage simplified 1-click operations from the UI or single REST API calls.

ROBIN Hyper-converged Kubernetes Platform

ROBIN offers a self-service app-store experience that simplifies deployment and lifecycle management and thus shortens DevOps and IT tasks from hours and weeks to minutes. Hyper-converged Kubernetes platform makes applications truly agnostic of infrastructure choices. This results in applications sharing resources and data with predictable performance, leading to significant cost savings.

With end-to-end resource management, ROBIN lets users control QoS and performance SLAs, dynamically adjust and guarantee resources. In addition, users enjoy 1-click lifecycle management operations to scale, migrate, snapshot, upgrade and patch even big data, databases and AI/ML applications. Most importantly, with ROBIN users can do self-service deployment of big data, databases and AI/ML. Also, teams can share entire experiments among team members and quickly do what-if trials. They can scale resources including GPU and IOPs. Finally, users can migrate as well as recreate entire application environments across data centers and clouds.

To learn more about hyper-converged Kubernetes platform architecture in detail and the philosophy behind it, download this white paper.


Author Partha Seetala, Founder and CTO

More posts by Partha Seetala, Founder and CTO