Hedvig Inc. laid out a vision for a Universal Data Plane spanning public and private clouds, as it launched an...
updated version of its software-defined storage product.
The Hedvig Distributed Storage Platform, which launched last year, can run on commodity hardware across multiple sites, either at a customer's data center or in public clouds. The Hedvig storage software clusters servers into a virtual pool of capacity and supports block (iSCSI), file (NFS) and object (Amazon Simple Storage Service, or S3, and OpenStack Swift) interfaces.
Version 1.0 included storage features such as deduplication, compression, snapshots, clones, tiering and caching. Version 2.0, due in October, will add capabilities tailored for multiple workloads, clouds and storage tiers.
Plug-ins for VMware vSphere Web Client, Docker volume and Mirantis' distribution of Fuel, the OpenStack management software, will enable customers to manage all aspects of Hedvig storage natively from inside those environments, according to Rob Whiteley, vice president of marketing at Hedvig, based in Santa Clara, Calif.
Hedvig also did the necessary integration work with Microsoft's Scale-Out File Server to provide an SMB version 3 share for users in Hyper-V environments. The Hedvig storage software does not natively support SMBv3, Whiteley noted.
Hedvig's cloud interoperability extends to Amazon Web Services (AWS), Microsoft Azure and Google Cloud Platform. For instance, the Hedvig software could run on Amazon Elastic Compute Cloud and use Amazon Elastic Block Store for storage. Hedvig would virtualize EBS and consume the underlying EBS volumes, Whiteley said.
"We're not using S3. We're not using cloud storage. We're instead using these public clouds almost as a hosted data center, allowing you to move applications or burst out to use those cloud services and guarantee the data is there, if need be," Whiteley said. "Technically, this was always possible in the past, but there were some new things we had to develop in order to make it easier to run out of the box. It's much easier now to spin up instances of our software in the public cloud."
Hedvig also did integration work for Apache Mesos and Google Kubernetes for cloud orchestration, and it introduced user-defined tagging to let customers generate and group virtual disks for management across multiple sites. Other new features include role-based access control for multi-tenant environments and an Ansible-based installer to ease multicloud operations.
Customers with multi-tier storage will find a newly added central management console to improve visibility across a Hedvig software deployment and beacon information back to Hedvig for predictive storage analytics. Hedvig is also adding storage virtualization for third-party arrays and cloud volumes, as well as Hedvig Storage Proxy enhancements to improve deduplication and server-side caching.
Universal Data Plane vision
The new version 2.0 capabilities broadly align with Hedvig's Universal Data Plane vision of a single programmable data management layer that enables any application workload to store and protect data across any site, including public or private clouds.
Capabilities Hedvig cited as crucial to the Universal Data Plane vision are a scale-out, software-based architecture; multisite replication to eliminate any single point of failure; and automation and orchestration to ease provisioning and application-specific, policy-based data services.
Avinash Lakshmanco-founder and CEO, Hedvig
"We blur the lines between what a customer-run data center site is and what the public cloud is right now," said Avinash Lakshman, co-founder and CEO of Hedvig. "You could have clusters span any site. You could run our systems across multiple cloud vendors -- perhaps Google, AWS [or] Azure -- and keep data replicated across all of them seamlessly."
Lakshman said the Hedvig storage model stems from his background. He co-invented the Amazon Dynamo NoSQL database and created the Apache Cassandra distributed data systems while at Facebook.
"The Universal Data Plane is perhaps the only way to avoid large vendor lock-in as people make their journey into the public cloud," Lakshman said. "[But] you don't need to actually run in a public cloud to take advantage of this. You could be running your environment across multiple sites, and when one is ready to move to the public cloud, you could use the same platform."
Henry Baltazar, a research director for the storage channel at 451 Research, said storage buyers are more accustomed to distinct NAS and SAN products in siloed environments. But, as they try to cope with explosive data growth with flat to only moderately increasing budgets, "that's opening up an opportunity for people to start thinking differently," he said.
Baltazar said Hedvig's "very aggressive strategy" has a chance to succeed, but he envisioned greater receptivity with "the next generation of IT people" who are "growing up around cloud and more tied to the applications."
He said initial use cases include test and development, unstructured data, and secondary storage and analytics. Hedvig's competition includes Dell EMC, NetApp, Nexenta and Red Hat, according to Baltazar.
"There's going to be a need for mobility in and out of different clouds, both on premises and between clouds," said Jeff Kato, a senior storage analyst at Taneja Group Inc., in Hopkinton, Mass.
"That's really what [Hedvig's] Universal Data Plane architecture is supposed to provide. The idea is, unless you have that, it's going to be much harder to move a workload around from cloud to cloud."
List pricing for the Hedvig storage software is 60 cents per gigabyte, per year. The price includes all features and maintenance. Customers deploy their own hardware or work with partners to bundle the software and hardware.
Hedvig on list of startups to watch
Storage vendors throw support behind Docker
Assessing software-defined storage benefits