Shifting the Database Paradigm
I’ve used databases for a long time, but I’ve taken them for granted as things that worked well for what I wanted to do and didn’t require much further thought. I feel kind of stupid that it had never occurred to me to think about how dependent they were on the concept of rows – that is, it is the row that dominates the database, not the column.
But once I started to read about Michael Stonebreaker’s company, Vertica, it become so obvious – of course, if you make the column central to the database instead of the row, you open up all sorts of possibilities because the data in each column is uniform. Imagine a simple database of names and addresses – each row contains the name and address for an individual, but each column contains only one sort of data, like the zip code. Making a database column-centric rather than row-centric means that you have uniform data to store and access. Vertica claims that it is faster by 80-fold to access the data in a column-centric database, as well as offering major advantages in compression for storage.
Once you are told this concept, it seems so obvious in its potential advantages that it is a total no-brainer. Yet I probably would not have thought of it on my own. I wonder how many other paradigm shifts are out there waiting for someone to make the rest of us see how obvious they are. Vertica has raised 20+ million for its database idea. As they say, current relational databases have stayed the same for 30 years. What else could benefit from a 180 degree turn?