Run-length encoding (RLE) is a very simple form of data compression.
RLE is a useful compression scheme when the same data is repeated a lot in the database. It can be used for graphics where there is mostly a single color background and a few lines forming the image. It may also have use in numeric databases where there are a lot of 0's in a row in the data. If the data becomes more complicated the encoding scheme can actually increase the size of the file. It is a lossless scheme as none of the original data is lost.
It works best when there are runs of data, that is when the data is identical for many successive instances. The RLE form of this data would replace all of the instances of a particular data steam with a one instance of the data and the count of how many times it is repeated.