vFilo’s file and object capabilities complement DataCore’s SANsymphony block storage, and while their interfaces are separate at the moment, the plan is to integrate them.
Software-defined storage vendor DataCore has supplemented its flagship SANsymphony offering with vFilo, its next-generation distributed file and object storage virtualization technology. In addition to complementing SANsymphony’s block storage, it also shares core aspects of its technology, including its ability to work with existing hardware and its scale-out and virtualization capabilities.
vFilo is the latest stage in the aggressive growth strategy DataCore embarked on over a year ago, when co-founder and longtime CEO George Teixera was moved off to be Chairman of the Board and Dave Zabrowski brought in to replace him. While the company has been profitable the feeling was that their market presence was rather less than their technology warranted, and the goal was to change that. This July, DataCore announced changes that they said were tantamount to a relaunch of the company, with a new version of SANsymphony being joined by a new hyperconverged infrastructure appliance, a new cloud-based analytics service, and subscription-based licensing. Now, with vFilo, they are rethinking how file and object storage fit into their portfolio.
“We have had file and NAS systems for many years, but vFilo is a completely new product,” said Gerardo Dada, who was brought into the company as CMO at the same time as Zabrowski. “It’s a completely new product. Most file and NAS architectures out there are pretty outdated. This is next generation because it changes way people think about data.”
Dada said that DataCore started building this a year ago with a clear vision in mind.
“We believe that everything should be covered by one storage layer, that things are more consistent when they are done one way for an entire infrastructure,” he stated. “We believe we can do this for block, file and object.”
So why create a second product instead of expanding SANsymphony? Dada said this was simply a more efficient way to do it.
“Logical first thinking was adding it on top of SANsymphony. That would have been common, but it would have been bolting an interface on top of another system. Making the file and object storage separate from block storage was a decision that we had to make, which reflected the fact that file and object storage is often higher volume, and includes secondary as well as primary storage.”
A major architectural distinction of vFilo is its separation of data and metadata for storage purposes, while maintaining the links between them
“It’s a unique architecture that lets us separate metadata and the data, and lets the system be super high performance,” Dada said. “Most companies use third-party metadata management engines with metadata. With vFilo, the data and metadata are integrated in a way that is not dependent on physical location of the data, so that they are still tightly integrated while physically separated. This means that data can be moved to less expensive storage in the cloud, but the metadata is still in the global namespace so it looks like data is still there. It can be retrieved that way for eDiscovery – just not as fast as if it was there. No other system has this separation of data and metadata.”
The Linux-based vFilO technology is designed to non-disruptively assimilate third party files spread across multiple file servers, NAS appliances and object storage, assimilating those silos under that global namespace accessible via standard NFS, SMB and S3 protocols.
“Some file companies just add an object interface on top, but this is problematic because they aren’t as fast as a file system,” Dada stated. “vFilo non-disruptively assimilates existing filers. In a few minutes to an hour, it will have all the metadata and can be the unified front end for the system, ingesting data from all the other systems. We can fetch data through parallel NFS at incredibly fast speeds. It also ensures policies can be done at a very granular level, and are quicker and more efficient than volume level.”
vFilo comes with what Dada described as a very full set of data services.
“We even provide the ability to Undelete, which most competitors don’t do,” he said. Effectively, this adds the equivalent of a recycle bin. Other services include dedupe, compression, encryption, enterprise-class snapshots, clones, and replication.
“The use cases include the consolidation of the mix and match of different systems which became islands, so that data can be tiered and moved between them,” Dada said. “It also allows for the leveraging of cloud the same way. And it will support traditional use cases for high performance system to store unstructured data.”
One key component to be done remains, although it is on the roadmap. Today, because of those original design decisions, SANsymphony and vFilo are two separate products, which have two different interfaces.
“There are things you do in vFilo that requires metadata richness which mean that there is a different user interface than SANsymphony,” Dada said. “We are working on an update of DataCore Insight services that will ingest it all. Right now, it’s single pane of glass for block, and a separate one for file and object, but eventually it will just be one pane.”
vFilo supports NFS v4.2, NFS v3, SMBv3 and SMBv2 protocols as well as Kubernetes CSI for persistent volumes. It is priced per TB, with one price for active data and a lower price for inactive archives. The software will be generally available on November 20, 2019 through DataCore channel partners.
“vFilo will be a great tool for partners,” Dada stated. “Today they sell us for block and someone else for file or object. Now they can sell block, file and object from Datacore.”