My name is Mayank, and I co-founded Aster Data Systems with George and Tasso in 2005.
Shortly after incorporation, the three of us were eating lunch at a Chinese restaurant and out popped a fortune slip from a cookie reading:
Indeed. The Internet is changing the speed at which we communicate, processes are being automated to react and execute in the blink of an eye, and data is playing a key role in guiding execution. Analysis of data is moving front-and-center, breaking out of the passive world of warehousing and reporting, as applications create intelligent processes, and companies live-and-die by their ability to monetize their data.
A new set of applications are being written – or waiting in-the-wings to be written – that will leverage data to act smarter. Consider the rapid evolution of online advertising networks: in the past 5 years, we have seen a spate of successful companies carving out a niche for themselves in the market. Their differentiation? The unique ability to match advertising inventory with consumer segments. Their basis of differentiation? Data!
And yet, a majority of these advertising networks do not use databases for their optimizations! Google and Yahoo! have famously built their own platforms; so did the amazingly talented teams at Right Media, Kosmix and Revenue Science. Of course, these companies use databases: but only for reporting and billing purposes.
How did we get to this cross-road where data is being analyzed outside the database?
For too long, databases have been clunky, monolithic systems that are rigid and inflexible, locking up the data in architectures that are
1. Hard to query
2. Hard to scale
3. Hard to manage
Meanwhile, the landscape of applications around a database is changing, shifting away slowly but surely.
We will use this blog to outline our thoughts on this changing landscape, along with our experiences in building an analytics database and a company that participates in this change.