In a development that turns object storage into a natural extension of existing file systems, Caringo today unveiled Firefly for Swarm.
Tony Barbagallo, vice president of product for Caringo, says that FireFly for Swarm was designed to be compatible with Windows and NetApp file servers. Using it with Swarm allows organizations to store hundreds of petabytes of data that remains accessible to applications that were built on top of traditional file systems.
Storage administrators can set file- and directory-level policies based on multiple file attributes to automatically move file data from NetApp or Windows file servers to the Caringo Swarm environment. Files can either be copied into Swarm for data protection or they can be moved. In the latter case, a lightweight pointer is left on the file server that an application can use to invoke that file when needed. If a file on Swarm is accessed, it gets placed back into primary storage to ensure native file server performance.
By reducing dependency on file systems for secondary storage, Barbagallo says IT organizations can reduce the total cost storage in their environments by as much as 400 percent.
Data stored on Firefly for Swam can be accessed over HTTP or using an Amazon S3-compatible application programming interface (API). Data can also be encrypted at rest and in motion and because Firefly for Swarm captures metadata. Barbagallo says those files can be searched based on source file name, folder path, MIME type, attachment content disposition, date creation and modification, attributes, owner name and or source host.
One of the major challenges IT organizations face when trying to manage storage at scale is the limitations of existing file systems. To get around those limitations, IT organizations have been adopting object-based storage systems, but applications written for file systems haven’t been able to make use of object-based storage up until now without having to access a type of gateway appliance. The arrival of FireFly for Swarm means that IT organizations can now use commodity storage to store secondary and tertiary data inexpensively at scale without requiring existing applications to be rewritten to natively interface with an object-based storage system.