I just read about Phoenix project. People from Phoenix talks very well about it, the question is,,, they give you SQL and it's suppose pretty fast.. any case where is it better to use just HBase without Phoenix? it is in general faster than execute native scans and make your own coprocessor?
