Data Analytics Startup Lands MySpace As Early Adopter

Three-year-old Aster Data Systems is about to launch its flagship product, an analytics database that scales to hundreds of microprocessors. The Silicon Valley startup has an impressive customer, MySpace, that's apparently already using its new system.Aster Data Systems' announcement isn't until next week, so details are sketchy. In an e-mail in advance of that, the company says MySpace plans to use its technology to analyze user behavior and advertising opportunities and to "control information risk."

That's a tall order for a brand new data-analysis platform. If Aster Data Systems can help MySpace in any one of those areas, let alone all three, it would be an impressive start for the company. I hope to talk to MySpace to learn about how it's using the system and why it's going with a newcomer to the business intelligence market instead of an established player.

Aster Data Systems was founded in 2005 by three colleagues in Stanford University's Computer Science Department. As a Stanford grad student, CEO and co-founder Mayank Bawa developed algorithms for querying distributed systems, while CTO and co-founder Tasso Argyros researched large-scale data clusters. Chief scientist and co-founder George Candea's resume includes stints with Oracle, IBM Research, and Microsoft Research.

That's some pretty deep expertise, which is what's needed for a startup to succeed in a mature and consolidating BI market. In the company's blog, CTO Argyros lays out what he sees as three common but unsatisfactory approaches to performing analytics against large data stores: 1) scale up by adding processors to a DBMS; 2) deploy a DBMS across a massively parallel cluster; 3) write custom code. "All of these approaches are pretty far from solving the real problem," he writes.

Just how Aster Data Systems does things differently remains to be seen. The company's Web site offers clues to its technology, but details are lacking. It sounds like Aster's data analytics "solution" isn't an actual data warehouse or appliance, but software that runs queries outside of a centralized data warehouse.

There's no shortage of vendors coming into the data warehousing and BI market, from Dataupia's appliance, to Pentaho's open source platform, to Truviso's complex event processing software, to Hewlett-Packard's NeoView. I'll report on what makes Aster Data Systems unique once I get a chance to talk to the company.