Research Library

Hedvig Distributed Storage Platform – Product Brief

Last updated December 28th, 2018 for annual review and updates. 7-page review of Software Defined Storage product Hedvig Distributed Storage Platform. Covers highlights, potential, usage, and strengths/weaknesses.

Hedvig Distributed Storage Platform is a software-only solution comprised of the Hedvig Storage Proxy, Hedvig Storage Service and Hedvig APIs. The Storage Proxy runs as a guest VM or Docker container and delivers data services and a client-side caching layer to compute environments, including all four major hypervisors, plus OpenStack Cinder. The Hedvig APIs provide access to the Hedvig Distributed Storage Platform for cloud and application developers via a RESTful interface.

The Hedvig Storage Service runs on runs on standard hardware, both x86 and ARM-based servers. Like other Software-Based Storage solutions, the Hedvig Storage Service abstracts local storage (Flash, HDDs) into pooled, virtual volumes for use by applications running on these servers or on external hosts. The software can be installed as a VM or can run natively on a Linux server. The product supports multiple storage protocols including iSCSI, NFS, S3, Swift and OpenStack Cinder.

In the Hedvig Distributed Storage Platform the Storage Proxy is independent from the Storage Service. This enables users to scale storage separately from compute, in what the company calls a “hyper-scale” configuration, or to scale storage and compute together in a “hyper-converged” configuration.

The Hedvig cluster requires two nodes at a minimum, three is recommended, but has no specified upper limit, although a 1000-node cluster has been tested. Hedvig can address any storage type installed in or direct-connected to each node, with per node capacities governed by the connectivity limits of the hardware. Existing storage arrays can also be direct- or network-connected in JBOD configuration as a way to utilize legacy storage capacity. Typical node configurations support between 20 and 32TB, based on disk drive sizes.

Instead of using RAID, the Hedvig architecture parses data into blocks, which are spread around the cluster based on policies, or a best effort basis, to maintain resiliency during failures. Flash is used throughout the cluster for caching and the system can pin data into flash to make a volume look like it’s on an all-flash array.

Product Brief Includes:

  • Highlights
  • Usage and Deployment
  • Eval(u)Scale Table
  • Evaluator Group Comments
To Download, Contact Us about subscriptions or purchase
LINKEDIN

Forgot your password? Reset it here.