Nitpick from the article: it suggests kdb underpins most algorithmic trading systems. This is simply not true. I've worked in a number of banks and hedge funds, and most have their own home-brew storage and analysis engines (I've helped build a few myself).