De-duplication engine using any language

Create a de-duplication engine that acts as a file storage, retrieval, and handling system. It must take some files as inputs, take data from it in chunks of 8 byte size and store it in some efficient data structure of our choice. The data structure should be robust and must not store duplicate chunks. Instead, it has to make a reference to the original chunk that is repeated.

What would be the most efficient solution to this?

I have a feeling we can use maps to solve it but please give your opinion.