Albireo Software Development Kit (SDK)

Ready-Set-Code!

The SDK is the product of choice for vendors that have already implemented their indirection layer, but are looking to add fast, scalable, and resource efficient deduplication. It provides  duplicate identification to any platform through a C language library with full documentation and support. Suitable for block, file, and object architectures, Albireo SDK is the only drop-in deduplication index on the market.  Its patented deduplication index is uniquely optimized to work with HDD, flash, and Hybrid storage. The SDK has a multi-node design which allows it to scale performance and capacity linearly across multiple servers.

Impact

Available for on-premises, cloud, or hybrid deployments, Albireo SDK delivers 3X-30X data reduction through deduplicating the data associated with these common storage use cases:

SDK is the ideal embedded deduplication solution for products from leading:

Competitive Edge

timerMaximum performance: 200,000 operations / second / processor core

Albireo SDK, delivers the fastest deduplication in the market – millions of IOPS – so even the highest performing storage solutions can incorporate inline deduplication without compromise.

microchipSmallest footprint:  0.1 bytes per entry in RAM

The culmination of over 10 years of research and development, Albireo SDK is the only deduplication indexing technology able to effectively index 10 TB of 4 KB blocks in as little as 1 GB of memory. Other common indexing techniques in use today require as much as 100 times the memory of the Albireo SDK.

arrowsHighest Scalability:  640 Billion objects per server, 10 Trillion objects per grid

Albireo SDK supports any scale-up or scale-out architecture. Dedupe rate is a function of chunk size and scale. Albireo SDK is the only solution available that can scale to PBs of storage at a 4 KB chunk size – for the highest dedupe rates and ROI.

How It Integrates

For systems with an indirection layer and reference counting, SDK integration can be completed in a matter of weeks using the six C language API calls provided. The vendor’s code simply:

  1. Calls the SDK, providing new data and location information,
  2. Receives a callback for the matched items, and
  3. Merges duplicate blocks in the indirection layer and updates reference counts.