Santa Clara, Calif. — By now, most people realize this is the age of convergence in IT – especially as it applies to storage. We have converged infrastructure mixing storage, compute and networking; hyper-converged infrastructure integrating compute, storage and virtualization in one box, and converged secondary storage putting backup, DR, archiving, test/dev, copy and cloud data on one platform.
Now startup Hedvig is pushing a new kind of convergence – primary and secondary data together in one distributed platform.
Hedvig designed its software-defined storage as scale-out, multi-cloud primary storage. But the startup finds early customers sometimes use it as a backup data deduplication target running on x86 servers. Hedvig CEO Avinash Lakshman said Hedvig software can drive primary storage that requires no separate backup.
“One capability we can bring to the table naturally is, if Hedvig is chosen as a primary storage platform, then you don’t need to take backups at all,” Lakshman said during a press briefing at Hedvig’s headquarters this week. “You can take scheduled snapshots in your primary environment, and go back to any snapshot from your primary environment. Think of it as converged where you have primary and secondary storage built it. We also provide the capability of moving snapshots to the public cloud as they age.”
Old-school backup admins will tell you this violates a cardinal rule of data protection. “It used to be, ‘Thou shalt not put backup data on the same box as primary,’” said Eric Carter, Hedvig senior director of marketing. “But distributed systems are no longer the same box.”
Hedvig also positions itself as a good fit for dev/ops because it includes self-service APIs to program and integrate applications.
Hedvig claims its software can run any workload on any infrastructure and over any cloud.
“We have been multi-cloud even before that term was coined,” Lakshman said.
Hedvig software forms a universal data plane supporting block, file and object storage. It installs on x86 nodes and cloud instances and forms a scale-out storage cluster over multiple sites and private and public clouds. Its storage proxy presents virtual disks at the application layer, routes I/O to the storage cluster, enables local flash-optimized services, and includes APIs for plug-ins and direct application integration.
Lakshman, who created Casandra and helped create Amazon DynamoDB as a developer, founded Hedvig in2012. Hedvig 1.0 software started shipping in 2015, and Lakshman said the company still has less than 50 customers. However, it has a few large customers since joining Hewlett Packard Enterprise Complete Program last June, a few months after HPE participated in a $21.5 million funding round.
Lakshman said the HPE reseller deal “has been a shot in the arm for us. They walk us to the table for deals we never could be part of, with Fortune 100 companies. We have a least half a dozen of those customers now. All those companies are pivoting toward hybrid and multi-cloud.”