Red Hat has launched what it claims is the first open source hyper-converged infrastructure product on the market.
Red Hat Hyperconverged Infrastructure (HCI) combines a number of Red Hat software products that can be deployed on recommended server hardware.
The foundation is the Red Hat Enterprise Linux operating system (OS). This is host to Red Hat Virtualization, the company’s KVM-based hypervisor environment, while storage comes from Red Hat Gluster Storage. The whole is provisioned and configured using Red Hat’s Ansible management software.
The move is effectively a bundling together of existing products, said director of storage product marketing, Ross Turk.
“We have taken a deployment that a lot of customers are doing in the field,” he said. “But we’ve done the tests to ensure the components will co-reside and play together nicely.”
Besides being pre-tested and guaranteed to work together well, another advantage of Red Hat’s hyper-converged software is that “customers can self-support and make use of community support”, said Turk.
The company also points out that all components, from OS to storage and hypervisor, are from the same supplier.
Turk added: “A lot of hyper-converged vendors have a lot of pieces from different vendors. All elements of our HCI are our own, from the OS through the hypervisor and storage to the orchestration – a single vendor for the entire stack from top to bottom.”
Red Hat Hyperconverged Infrastructure is a software product, aimed at smaller deployments at remote offices and small and medium-sized enterprises (SMEs).
The cluster must be three, six or nine physical machines with adequate memory and storage space. Red Hat recommends that each host should have two, six or eight CPU core, and between 48TB and 80TB of storage.
Hyper-converged infrastructure has been a rising star in the datacentre in the past couple of years. It combines compute and storage with a hypervisor, often as a hardware bundle, but sometimes, as in this case, as a software product.
The key advantages of hyper-converged infrastructure are that it allows for relatively simple deployment of compute and storage capacity, usually tailored to a specific virtualisation environment. Scaling is also usually a simple matter of adding extra nodes in a grid-like fashion.