What Doug said but also, if cells are of a size and/or data type that is compressible (no point compressing small stuff, same for trying to compress images types), then there are big wins all around if you have the application compress the cells before putting them into hbase; network traffic is less, the size of the data that traverses the hbase server is less (and your caches' can carry more 'data'), etc.
stack On Sat, Jul 16, 2011 at 4:38 AM, Doug Meil <[email protected]> wrote: > > Hi there, see this in the book: > > http://hbase.apache.org/book.html#compression > > And this... > > http://hbase.apache.org/book.html#trouble.client.longpauseswithcompression > > And see this thread which was the original discussion on the long client > pause entry. > > http://search-hadoop.com/m/WUnLM6ojHm1/Long+client+pauses+with+compression& > subj=Long+client+pauses+with+compression > > > > On 7/16/11 5:23 AM, "Sam Seigal" <[email protected]> wrote: > >>Hi All, >> >>A quick question on compression. I saw that HBase can use LZO compression >>for storing data into the HFile. >> >>Has anyone done experiments with using compressions at the application >>level >>instead instead of letting HBase handle it ? Are there >>advantages/disadvantages of this approach ? >>Is it possible to use other compression libs in Hbase other than LZO ? >> >>Also - are there best practices and/or conventions on encrypting data >>before >>storing it in HBase ? >> >>Thanks, >> >>Sam > >
