Application Aware Storage For GKE and OpenShift
Summary
As more and more enterprises move to Kubernetes for their cloud-native needs, they are looking for high-performance storage options for their stateful and data-intensive workloads. Beyond a seamless workflow to provision storage, they are also looking for complete data lifecycle management features as a part of their storage platform. Robin Storage is a perfect candidate for this need. With their application awareness, high performance and scale, Robin Storage is well suited to meet enterprise cloud-native storage needs. Robin has already partnered with Google Cloud and Red Hat OpenShift to deliver their customers with high-performance storage. They are also well positioned to work with any Kubernetes platform.
Introduction
As Kubernetes gains increasing enterprise adoption, the need for supporting stateful applications becomes a critical need. Kubernetes community is working on supporting stateful applications on the platform but the enterprise needs are different from running basic stateful apps. Many enterprises are forced to migrate their mission-critical legacy applications to container environments to ensure that both their legacy and modern applications are hosted on Kubernetes clusters. This brings into focus the need for a comprehensive storage platform that meets the critical enterprise needs such as:
- Ease of use without developers having to manage additional operational overhead.
- Provisioning and managing storage should be integrated into the developer workflow
- Bare-metal like performance
- High availability
- Comprehensive data management capabilities
While most Kubernetes platforms meet the compute and networking needs of enterprise workloads, very few platforms are capable of meeting the performance and data management needs of the enterprises. According to Rishidot Research’s analysis of market trends and our discussions with enterprise stakeholders, more than half of the enterprises are expected to deploy container based workloads by 2020. Almost three-thirds of our research participants are planning to use containerized workloads also said they are planning to deploy Stateful workloads on Kubernetes. There is a clear market need for storage solutions that work seamlessly with Kubernetes. As more and more big data and machine learning workloads move to Kubernetes, high performance becomes the core requirement for Kubernetes storage. Robin’s platform features meet these requirements and, recently, they announced partnerships with both Google Cloud and Red Hat OpenShift. In this short whitepaper, we will talk about the state of stateful applications in Kubernetes and how Robin Storage meets the enterprise needs of Google Cloud and OpenShift customers.
The state of stateful applications in Kubernetes
Running stateful applications in Kubernetes is challenging even with the progress made by the Kubernetes community on StatefulSets and support for many Persistent Volumes. According to the CNCF survey, the critical challenges for running containers in production are security, complexity, storage, and networking. The enterprise needs are much different from modern applications as they have legacy and data-intensive applications in the mix. Even with StatefulSets, trying to bring up clusters, handling node failures, managing routing, ensuring HA, etc. for databases or big data workloads requires a great deal of coding and operational overhead. Using storage volumes with Kubernetes includes managing volume level primitives that adds operational overhead. As developers push for ease of access to resources, managing multiple configuration settings to deploy stateful applications is a non-starter for developers in the modern enterprise.