Compression Algorithm

The book Inside the Windows NT File System provides a high-level description of NTFS compression.

Basically, the NTFS file system provides real-time access to a compressed file, uncompressing the file when it is opened and only compressing when it is closed.

When writing a compressed file, the system reserves disk space for the uncompressed size. The system gets back unused space as each individual compression buffer gets compressed.

Note

Some programs do not allocate space before beginning the save, and only pop up an error message when they run out of disk space.

NTFS compression uses a 3-byte minimum search rather that the two-byte minimum used by DoubleSpace. This type of search enables a much faster compression and uncompression (roughly two times faster), while only sacrificing two percent compression for the average text file.

Each NTFS data stream contains information that indicates if any part of the stream is compressed. Individual compressed buffers are identified by "holes" following them in the information stored for that stream. If there is a hole, then the NTFS file system knows to uncompress the preceding buffer to fill the hole.