Benefits

Major Advance Over Previous Delta Encoding & Binary Differencing

Finds matches without metadata or pointers identifying data as “derived from” prior data sets. Results in a dramatic increase in matches found, leading to substantially higher efficiency of replication and transfer.

Massive Parallel Processing

Made available by the cloud.  Hundreds of compute nodes deployed for a few seconds to detect matches.  Difficult prior to the cloud and its horizontal scalability.

Private Data Centers

Performance not quite as high as in the cloud, due to lack of scale, but still provides significant efficiency gains.

Process

1. The Ilesfay CDE™ (Content Delivery Engine) instructs each location to create digital fingerprints for all of the data files stored at each location that are part of the replication scheme. These “locations” can be servers at private data centers, cloud persistent storage, or a combination of both.  The digital fingerprints identify small parts of each binary object that make up the data files.
2. The digital fingerprints are all stored in the CDE and are labeled by location.
3. When new data files are created at a location, digital fingerprints for these are generated and sent to the CDE.
4. The CDE then compares the digital fingerprints for the newly created data files against all previous digital fingerprints.
5. To replicate the data files, the CDE uses the knowledge of which digital fingerprints match to create a patch file.  This patch file contains only the differences between the data files along with instructions on how to reconstruct the entire file.
6. The patch files are then transmitted by the CDE to each of the locations to complete the replication.

NOTE: For multi-location replication, various locations will likely required different patch files to be generated by the CDE.