Array Cow

Introduction

In memory array de-duplication, useful for efficiently storing many versions of data.

This is suitable for storing undo history for example, and is effective with both binary and text data.

• Configurable block sizes.
• Supports array-stride to avoids overhead of detecting blocks and un-aliened offsets. (a stride of 1 for bytes works too)
• Uses block hashing for de-duplication.
• Each state must only reference its previous, making both linear and tree structures possible.

Further Work

It may be worth using mmap for data storage.