Scality Extends Object Storage Reach to Kubernetes Clusters
Scality has extended the reach of its object storage platform to add support for the Container Object Storage Interface (COSI) and the Container Storage Interface (CSI).
COSI is a Kubernetes standard that enables cloud-native applications to dynamically provision and consume object storage, such as S3 buckets, using workflows based on Kubernetes application programming interfaces (APIs). The implementation being made available by Scality automates bucket provisioning, credential management, and access controls to eliminate the need for manual configurations or custom scripts. Application developers can, for example, define bucket claims and access resources from within their Kubernetes manifests in seconds.
The CSI driver, meanwhile, enables native integration with Kubernetes clusters, allowing IT teams to provision, attach, mount, and manage file storage volumes directly. It also enables file-based access for Kubernetes workloads that require access to a POSIX-like file system volume.
Both open source drivers are being made available on RING, an object storage platform designed for organizations that are deploying monolithic and cloud-native applications that need to access massive amounts of shared data, and ARTESCA, a set of pre-configured object storage systems that are designed to be deployed via a graphical user interface (GUI) by IT administrators who don’t have a lot of storage management expertise.
Rob Orton, senior technical marketing manager for Scality, said that approach enables IT teams to support a wide range of multi-tenant applications more cost effectively.
It’s not clear how many stateful applications are now being built and deployed on Kubernetes clusters. Stateless applications still outnumber stateful applications, but as more databases are deployed on Kubernetes, the need to attach storage directly to these platforms has steadily increased.
A recent Cloud Native Computing Foundation (CNCF) survey noted that nearly three quarters (74%) of organizations reported using containers in stateful applications, but the number of companies planning and/or evaluating future use of containers for stateful applications is only 44%.
At the same time, it’s worth noting that as the number of artificial intelligence (AI) workloads destined for Kubernetes clusters increases, the number of stateful applications should rise considerably in the months and years ahead.
Rather than requiring organizations to acquire a separate storage platform for cloud-native applications running on those clusters, Scality is essentially making a case for consolidating data on a storage platform that is accessible to both Kubernetes clusters and virtual machines hosting legacy monolithic applications. The overall goal is to provide IT teams with greater agility regardless of what type of stateful application is running, especially as more organizations choose to repatriate some applications running in the cloud to on-premises IT environments, noted Orton.
The one thing that is certain is that most organizations will be running a mix of stateful and stateless applications for many years to come. The challenge and the opportunity now is to determine how best to make data available to any type of application because even a stateless application will eventually need to access data residing somewhere. The only question that remains is how much it will cost to deliver the required level of performance.