Data compression algorithms for TinyOS.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
scripts
tos
.gitignore
README

README

README for tinyPack
================================================================================
Description

This is a library of data compression algorithms for TinyOS. The implemented
algorithms can compress arrays of up to 256 bytes at a time. To make
compression suitable to low-end embedded systems, the primary objective is to
reduce the RAM usage. None of these implementations use more than a few hundred
bytes of RAM. The second most important object is speed of compression and
decompression. Popular compression algorithms such as LZW, typically require
several kilobytes of RAM to achieve useful compression, making them unsuitable.

The code is still in the experimental stages. Here is a brief introduction to
the compression algorithms available in tos/lib/tinypack. PC based
decompression is available in scripts/tinypack/tinypack.py.

LzrwEgaChainC
	512 Bytes RAM, Fast, Good compression ratio
	This is an LZRW-like compression algorithm, which means that is uses a
	hash table to find repeat sequences of bytes that can be compressed
	down. LZRW encoding is similar to LZSS, but the hash table method makes
	compression fast, at the cost of a 128 byte table in RAM. Bytes that are
	not part of a repeated sequence, are encoding individually using an EGA
	(Exponential Golomb Adaptive) codebook. This codebook is fast, achieves
	good compression when bytes fit the exponential model, and uses 384
	bytes of RAM.

LzrwChainC
	128 Bytes RAM, Fast, Moderate compression ratio
	Same as LzrwEgaChainC, except no codebook is used to encode literal
	bytes. This uses less RAM, but reduces the compression ratio.

LzssChainC
	No RAM, Slow, Moderate compression ratio
	This is an LZSS-like compression algorithm. The advantage of LZSS is
	that is can be implemented without using any RAM. This is because
	the search for repeated sequences is performed iteratively over the
	previously compressed data. The cost is that this search is slow.

The TinyOS code is compatible with TinyOS 2.x.
================================================================================
License

For licensing details, see individual file headers.
Licensed under the Apache License, Version 2.0.