fotohansel - Fotolia

Scality launches Artesca container-native object storage

Object storage specialist gets on a burgeoning container object storage market with Artesca, which runs from Kubernetes to provide persistent storage for containerised applications

Scality has launched Artesca, which will provide persistent object storage for Kubernetes that runs from the container orchestrator cluster. It can provision object storage to container instances across on-site datacentres and the cloud.

Artesca will also hook up with HPE for a six-month period of exclusivity in which it will be available on all-flash and hybrid flash Apollo and Proliant servers, as well as as-a-service via Greenlake. It is a software product, so will be available to be deployed on any compatible compute hardware.

According to Scality co-founder and field CTO Brad King, Artesca is a fully containerised software stack and all services are packaged and executed as containers.

In the first release, to enable deployment on bare-metal servers – from HPE initially – Artesca comes pre-packaged with Scality’s MetalK8s Kubernetes distribution.

“Over time, this will be expanded to support emerging third-party Kubernetes distributions,” said King. “We are currently considering environments such as HPE Ezmeral Container Platform, VMware Tanzu, and Red Hat OpenShift.”

Artesca uses the not-yet-ratified container object storage interface (COSI) to interact with storage instances.

COSI is the object storage equivalent of CSI, and brings the ability to describe object storage for use as persistent storage in Kubernetes container deployments.

“We expect that COSI will come to a meaningful 1.0 specification in the coming months, to allow us to implement a COSI provisioner for Kubernetes,” said King.

Object storage for containers is a new frontier in the market, with startup Min.IO recently claiming to be first to market in the space.

Read more about container storage

Artesca is aimed at enterprise-scale use cases, but can be deployed in a single server. At scale, it can allow multi-tenancy, enterprise security and access functionality, with data protection from Scality’s erasure coding schemas. Even on a single node, Artesca can be rebuilt using Scality’s local repair codes, as long as there are at least 10 drives deployed.

Target audiences are application owners and DevOps engineers looking for fast, S3-compatible object storage. Scality has in mind edge uses cases such as internet of things (IoT) and data collection, analytics, medical imaging and media delivery.

Backup is an increasing area of use for object storage and here also, Scality hopes ISVs and suppliers will incorporate Artesca into products they develop.

Object storage suits use cases where data is unlikely to be modified, but is rather kept for future use. That is in contrast to file and block storage, where parts of files can be changed a block at a time. With object storage, entire objects are updated by replacement via Rest APIs.

Artesca shares significant amounts of code with Scality’s Zenko, which is the company’s “multicloud data controller”. It allows customers to move, replicate, tier, migrate and search data across on-premise, private cloud locations and public cloud.

Zenko is based on Scality’s 2016 launch of its S3 server, which provided S3 access to Scality Ring object storage. The key concept behind Zenko is to allow customers to mix and match Scality storage with capacity from different cloud providers.

Last September, Scality added the ability to use all-flash servers in the upgrade to version 8 of its Ring software. The move came amid a growing trend for object storage to move to more high-performance hardware with demand for analytics, IoT and content delivery use cases. It is also driven by the need to support applications that demand a higher-performance tier for use in certain workloads, such as recovery in some backup applications.

Previously, Scality only supported the use of a small amount of flash in its deployments.

Read more on DevOps

Data Center
Data Management