What is GZ (Gzip)?

What is GZ (Gzip)?

GZ (Gzip) is a file compression format and software program created by Jean-loup Gailly and Mark Adler in 1992. It uses the DEFLATE compression algorithm, which combines LZ77 lossless data compression with Huffman coding. Gzip was created as a free replacement for the Unix compress program, which used an algorithm encumbered by patents. The .gz extension indicates a single file compressed with gzip; for directory archives, gzip is typically combined with tar to create .tar.gz or .tgz files.

Gzip is the standard compression tool on Linux and Unix systems and is built into virtually every Linux distribution. It is widely used for compressing log files, configuration files, and software packages. Web servers use gzip compression to reduce the size of HTML, CSS, JavaScript, and JSON responses before sending them to browsers, typically reducing text content by 60 to 80%. The gzip format includes a simple header with filename and modification time, followed by the compressed data stream.

How GZ Works

A .gz file is the output of gzip, a single-file stream compressor. The key thing to understand is that gzip compresses one stream of bytes and nothing more: it does not bundle multiple files, it does not store a directory tree, and it keeps no concept of folders. When you see something like backup.tar.gz, the bundling was done first by tar and gzip was then applied to the single resulting stream. This clean separation of jobs, one tool to archive and another to compress, is a defining Unix design choice.

Internally gzip uses the DEFLATE algorithm, which combines two classic techniques. The first is LZ77, a dictionary method that scans the data for sequences it has seen recently and replaces each repeat with a short back-reference, a pointer saying copy N bytes from this many bytes ago. The second is Huffman coding, which then assigns shorter bit patterns to the most frequent symbols and longer ones to rare symbols, squeezing out the remaining redundancy.

The file itself is a small header, the DEFLATE-compressed payload, and a trailer. The header records the original filename and a modification timestamp, and the trailer stores a CRC-32 checksum and the original uncompressed size. That checksum lets a decompressor confirm the data came back exactly as it went in, which is why gzip is considered safe for long-term storage of the streams it handles.

gzip exposes compression levels from 1 (fastest, least compression) to 9 (slowest, smallest). Because DEFLATE decompresses very quickly regardless of the level chosen at encode time, gzip is popular anywhere data is compressed once and read many times, such as web servers sending gzip-encoded responses to browsers on the fly.

History and Standardization

gzip was written by Jean-loup Gailly and Mark Adler in the early 1990s as a free replacement for the older compress utility, whose LZW algorithm was encumbered by patents at the time. Being unencumbered and freely redistributable, gzip was adopted by the GNU project and spread quickly across every Unix-like system.

The formats it depends on are openly documented. The gzip file format is specified in RFC 1952, and the underlying DEFLATE algorithm in RFC 1951. Because these are stable, published specifications rather than a single proprietary program, gzip decompression is built into countless libraries, most notably zlib, the same core code that powers PNG images and much of the web.

Decades later gzip remains a default. It is the baseline content-encoding for HTTP, a standard step in software distribution on Linux, and the g in the ubiquitous tar czf command. Newer algorithms compress better, but gzip's combination of speed, tiny memory footprint, and universal support keeps it in daily use.

GZ Compared to Bzip2, XZ, and Zstandard

Against bzip2 and XZ, gzip trades ratio for speed. bzip2 (a .bz2 file) usually produces a noticeably smaller result but compresses and decompresses more slowly; XZ, built on LZMA2, compresses smaller still but is slower and needs much more memory. gzip is the fast option that wins when you value throughput over the last few percent of size.

The modern challenger is Zstandard (zstd), which at its default settings matches or beats gzip's ratio while decompressing several times faster. Where zstd is available it is often the better technical choice, and it has been added to many Linux distributions and web servers. gzip's advantage over it is purely historical reach: essentially every device, language, and tool already speaks gzip.

It is worth repeating that gzip is a compressor, not an archiver. To keep permissions, ownership, and a folder hierarchy you must pair it with tar, producing a .tar.gz. A bare .gz is only ever the compressed form of one file, which is exactly why it is so common for single large logs, database dumps, and web assets.

Technical Details

GZ vs Other Archive Formats

GZ compresses one stream quickly, trading some ratio against the slower but tighter BZ2 and XZ.

Pros and Cons

When to Use GZ (Gzip)

Here are the most common situations where GZ (Gzip) is the right choice: