[HBASE-4218] Data Block Encoding of KeyValues (aka delta encoding / prefix compression

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.94.0
• Fix Version/s: 0.94.0
• Component/s: io
• Labels: compression
• Hadoop Flags: Reviewed

Release Note

Adds a block compression that stores the diff from the previous key only. Good for big keys and small value datasets. Makes writing and scanning slower but because the blocks compressed with this feature stay compressed when in memory up in the block cache, more data is cached. Off by default (DATA_BLOCK_ENCODING=NONE on column descriptor). To enable, set DATA_BLOCK_ENCODING to PREFIX, DIFF or FAST_DIFF on the column descriptor. Set ENCODE_ON_DISK to true on column descriptor to have the encoding in place out in the hfile (on by default).

• Assignee: Mikhail Gryzykhin
• Reporter: Jacek Migdal
• Created: 17/Aug/11 19:24
• Updated: 27/Aug/15 07:51
• Resolved: 11/Feb/12 05:50

Description

A compression for keys. Keys are sorted in HFile and they are usually very similar. Because of that, it is possible to design better compression than general purpose algorithms,

It is an additional step designed to be used in memory. It aims to save memory in cache as well as speeding seeks within HFileBlocks. It should improve performance a lot, if key lengths are larger than value lengths. For example, it makes a lot of sense to use it when value is a counter.

Initial tests on real data (key length = ~ 90 bytes , value length = 8 bytes) shows that I could achieve decent level of compression:
key compression ratio: 92%
total compression ratio: 85%
LZO on the same data: 85%
LZO after delta encoding: 91%
While having much better performance (20-80% faster decompression ratio than LZO). Moreover, it should allow far more efficient seeking which should improve performance a bit.

It seems that a simple compression algorithms are good enough. Most of the savings are due to prefix compression, int128 encoding, timestamp diffs and bitfields to avoid duplication. That way, comparisons of compressed data can be much faster than a byte comparator (thanks to prefix compression and bitfields).

In order to implement it in HBase two important changes in design will be needed:
-solidify interface to HFileBlock / HFileReader Scanner to provide seeking and iterating; access to uncompressed buffer in HFileBlock will have bad performance
-extend comparators to support comparison assuming that N first bytes are equal (or some fields are equal)

Link to a discussion about something similar:
http://search-hadoop.com/m/5aqGXJEnaD1/hbase+windows&subj=Re+prefix+compression

Issue Links

• relates to: HBASE-14323 Encoding rpc payload instead of compression