Ocarina Networks seeks to provide compression specifically for the data already in your online storage media. Their two products--the Ocarina Optimizer and Ocarina Reader--work together to both optimize the existing data regardless of storage media or vendor as well as provide real-time access to that data, for both users and applications, after it has been compressed.

At the core of the Ocarina product line is "file-aware" technology that facilitates the optimized compression of data based on detection and specific knowledge of the types of files that are being compressed and the applications that initially created them. The Optimizers read the existing file-based stores, identify each file, and apply a three-step "ECO" process (Extract, Correlate, and Optimize) to their reductions: First, the raw storage objects within the file are specifically extracted (such as document de-layering, de-compression of already compressed data, etc.). Next, these extracted components are correlated both within the same file and across multiple files; a process that the vendor states is able to identify and remove both exact data matches as in traditional de-duplication schemes as well as matching "similar" storage objects based on the type of information being processed. Finally, the remaining storage objects are optimized with specific algorithms applied based on the type of file data being crunched. The vendor states that their optimization efforts result in compression ranges of 40% for complex image data to 90% for "common office file mixes."

The optimization itself is performed out-of-band, via the Ocarina Optimizer. The Ocarina Optimizer is a pre-loaded hardware appliance that is deployed on the network and acts as an independent NAS client, reading files from existing storage after they've already been initially written by the original applications (filters can be defined governing which types of files should be processed, by location, age or type) and writing back optimized files as a result. Because of their out-of-band operation, complete failure of the Optimizer would result only in the storage data not being compressed. The initial Optimizer in the product family is the 4400; which boasts sixteen 64-bit processing cores and, when deployed in pairs, can process as much as 5 TB of data a day. Both the Optimizers and the Readers can be installed in load-balanced clusters as required.

Working with the files compressed by the Ocarina Optimizer is the job of the Ocarina Reader, and thus a complete Ocarina platform must include at least one Optimizer and one Reader. The Reader itself can be purchased as software (can be deployed as a file system filter driver on Linux-based Web/application servers, or on a dedicated server) or the 2400R hardware appliance, deployed between the NAS storage and its clients. The Reader provides access to the files (read and write) compressed by the optimizer in real-time, with the vendor noting that the compressed files themselves are stateless in nature (they can be copied from media to media and read by any Ocarina Reader in any location). Both NFS and WebDAV traffic streams are supported by the Reader.

New features of the Ocarina product platform focus both on the application of the compression technologies specifically to tiered storage deployments, and expanded file awareness for industry specific file types. For tiered storage, the vendor has introduced ECOmove, which facilitates one-step, policy-based migration and optimization of files from the primary to the archived (or optimized) storage tier; ECOsnap, which facilitates the taking of daily snapshots of storage, snapshots that are themselves optimized/deduped based on previous days' snapshots; and support for virtual global namespace technology, where files can be physically moved from location to location without effecting the virtual location from which they are accessed by end users. Expanded file awareness includes those for media post-production (RLA, AVI, Maya, EXR, Renderman, SPM, more), the oil and gas industry (online seismic data archives, including compressors for such industry specific files as seg-y, seg-b, and BRI), and medical image archives (for images in X-rays, MRIs, PET scans, gene sequencing data, and other files).

The Ocarina Plaform is available now. Visit the vendor's Web site for further information.