Artur Marciniec - Fotolia

News Stay informed about the latest enterprise technology news and product updates.

Caringo Swarm refines migration with FileFly software

New FileFly software for Caringo Swarm lets Windows Storage Server and NetApp filers migrate data to Caringo object storage systems.

Caringo this week came out with software that lets customers move data back and forth between file-based primary storage and Caringo Swarm object storage software.

FileFly for Caringo Swarm is a Windows-based application that plugs directly into the Windows NTFS file system. FileFly uses policy-based automation to identify and migrate aged data from primary NetApp file servers and arrays running the Windows Storage Server operating system to Caringo Swarm on the back end. No changes are required to applications or end user workflows.

FileFly for Swarm works with version 7.5.2 of Swarm, which the vendor released in July.

FileFly replaces Caringo Content File Server, a Linux-based archiving tool that provided an NFS mount on top of Swarm. Content File Server required storage admins to have some familiarity with Linux storage.

"FileFly is designed to more closely match the performance of native file servers," said Adrian Herrera, Caringo vice president of marketing.

Swarm presents data as objects. The vendor claims the combination of Swarm and FileFly helps reduce primary storage costs and retain the performance of standard file storage.

"Filers are becoming junk drawers filled with data that hasn't been accessed in months or years, but no one wants to delete it," said Tony Barbagallo, Caringo vice president of product.

"That's the specific use case we set out to solve: an automated way to transparently archive data from primary storage into lower-cost secondary storage," he added.

That's the specific use case we set out to solve: an automated way to transparently archive data from primary storage into lower-cost secondary storage.
Tony Barbagallo, vice president of product at Caringo

Caringo Swarm scales to support more than 100 PB of object storage. Swarm evolved out of Caringo's CAStor, a content-addressed storage system.

Unlike file and block storage, object storage uses a flat address space to store files, images and data blocks as individual objects. Each object is assigned a unique identifier for retrieving data from any location. Object storage software uses erasure coding for data protection.

Adoption of object storage has been slowed by the labor-intensive process of migrating metadata from legacy storage systems. Marc Staimer, president of Dragon Slayer Consulting in Beaverton, Ore., said FileFly for Swarm should help Caringo narrow the gap with object storage leaders Cleversafe and Scality.

"It's always a pain to move data," Staimer said. "The biggest barrier to moving data to object stores is relinking them and maintaining chain of ownership. Caringo is doing it in an automated fashion that maintains chain of ownership. That's pretty clever, and it should help them sell more object storage."

Caringo said it does not plan to bundle FileFly in upgrades of Swarm object software. Swarm customers can download FileFly software directly for Windows and NetApp file storage from Caringo. The vendor is making available a FileFly starter kit for new customers that scales to 140 TB of raw storage (100 TB usable) for $35,000.

Next Steps

Object and software-defined storage: When to use the popular technologies

Object storage advantages over file storage

Object storage software use cases include archiving, backup


Dig Deeper on Storage migration

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Which obstacles need to be addressed before you adopt object storage software?
Some customers need object storage, and that's fine. But the LAST thing a company with a "junk drawer" filer needs is to stub that data off to a backend object store. All this does is delay the day of reckoning when the drawer must be cleaned, and makes the eventual process much more painful because of throughput bottlenecks caused by the tiering layer in a performance-intensive NAS migration scenario.

It also makes a Windows server environment more fragile, as the filesystem driver which handles recall is certainly less mature than the native Windows filesystem stack (which has billions of hours of customer runtime at this point). Adding latency and reducing reliability is not generally an improvement in most storage environments, even if this software allows a single Windows Server instance to address "100 PB" of storage.
Well, I'm not opposed to policy based software to migrate data from primary storage to secondary object storage, but installing an application that "plugs itself" into the NTFS file system does raise a cautionary flag. Hierarchical Storage Management or HSM is replete with disappointment, and its policy-driven reincarnation as Information Lifecycle Management may be no better. It might be more useful if some intelligent analysis was brought to bear on the data being subject to this process. Something like what DataGravity is doing comes to mind, which would be even more useful, if it could tier data to S3-compatible object storage. No one is arguing for keeping useless files around, but first you have to know whether it really is useless or if it should be better protected than it currently is because it contains sensitive or proprietary information.
In addition to Caringo, SwiftStack added version 3, which also lets customers access file and object storage with S3, OpenStack and REST-based APIs, as well as standard file system protocols.