Amidst all of the controversies and atrocities plaguing the world right now, Kubernetes has carried on trailblazing the ever evolving technology landscape. As a result, customers are running more and more stateful workloads on Kubernetes clusters, Rancher Labs CEO Sheng Liang told SDxCentral. 

More data means more storage, and by 2025, IDC expects worldwide data to grow by 61% to 175 zettabytes. As organizations continue making the shift to Kubernetes, DevOps teams are fumbling with the management and storage of data rich or stateful applications. 

Rancher Labs responded to the DevOps distress call last week announcing the general availability of Longhorn, its cloud-native container storage service that enables vendor-neutral persistent storage that supports stateful workloads on Kubernetes clusters. The beta version of Longhorn launched in 2017, and was donated to the Cloud Native Computing Foundation (CNCF) as a sandbox project in 2019. 

Since Kubernetes already features a standard plugin mechanism by way of a container storage interface (CSI) that connects Kubernetes clusters to different types of persistent storage implementations, Liang explained, the concern is not so much in data portability but in the exact level of support that often differs from cluster to cluster.

“Longhorn complements Kubernetes,” Liang said in explaining how Longhorn provides a storage overlay that virtualizes and enhances underlying storage infrastructure. 

Rancher Labs' Longhorn is a “portable persistent storage services layer that can be distributed with a stateful application so that the application no longer depends on the level of persistent storage support provided by the underlying Kubernetes cluster,” he added. 

And while containers have become the next battleground for infrastructure-as-a-service (IaaS) providers – one where Amazon Web Services (AWS), Microsoft Azure (Azure), and Google Cloud Platform (GCP) each offer compelling services that further break down barriers of entry to the cloud – Liang said that hosted Kubernetes services like AWS’ Elastic Kubernetes Service (EKS) and Azure Kubernetes Service (AKS) offer built-in persistent volume support, but miss the mark with on-premises or edge Kubernetes clusters.

“As a result, there’s no guarantee a Kubernetes cluster can provide the persistent storage support needed by a given stateful workload,” Liang said. “In addition, there is no standard way to migrate persistent storage data from one Kubernetes cluster to another.” 

Storage has been one of the more difficult areas for Kubernetes to proliferate for the fear of losing a piece of data — and understandably so, as containers have continuously missed the mark on persistent data. Each container acts as a physical server or virtual machine (VM) where all applications have a little piece of data. This eliminates the opportunity to manage a single storage volume. 

Gartner predicts that 75% of global organizations will be running containerized applications in production by 2022. “This rapid uptake of containerized ecosystems has exposed a gap in the storage industry, where traditional storage providers are unable to provide functionality and support,” wrote Wendie Cheung, senior manager of global product marketing at Rancher Labs, in a blog post. 

Liang expects the need for cloud services and demand for cloud-native technologies to continue to grow in the months ahead as it is more important for organizations to “expedite the development of cloud services that are built using cloud-native technologies and run on Kubernetes clusters.”