Tuesday, June 15, 2010

Columnar Databases and the Future of Analytics

Delivering deeper data analysis, in formats more consistent with end-user expectations, has been a passion at Jaspersoft for several years. This passion will continue because making analytics, up to and including the sophisticated multi-dimensional type, more accessible and usable must be the goal of the BI provider who claims “pervasiveness” as its mantle. To help our community and customers know and influence where we’re going, we’ve published white papers that describe our view of the broader analytics landscape as well as one that describes the different types of analysis that our tools offer (e.g., in-memory analysis versus traditional, OLAP analysis).

Making data analysis a more common part of every business person’s day is fast becoming the province of both the database and the analytic tool. To advance the analytic agenda, columnar data storage engines, complete with speedy, in-memory techniques, help streamline query processing - delivering results in an instant and at a very low cost. Jaspersoft and its customers have been taking advantage of this trend growingly during the past two years. For example, in addition to analytic databases for enterprise data warehouses like GreenPlum, Jaspersoft has certified its products on columnar analytic databases including Vertica and Infobright, and we have many joint customers together. In many cases, the combination of in-memory-based analytics (within our BI tool) and the strengths and speed of the analytic database, obviate the need for building and maintaining not only star schemas in the data base but traditional OLAP cubes for the BI tool as well. Now that’s the beginning of analytics for everyone.

Interestingly, the pioneer in columnar technology for business intelligence usage, Sybase IQ, is one of the gems to be picked up by SAP. It recently announced its intention to acquire Sybase, and most speculation around the intentions have focused on Sybase’s treasure chest of mobile solutions. While I appreciate the mobile aspects, the combination of SAP’s Business Objects with Sybase IQ could be a formidable, albeit expensive and proprietary, competitor in the changing analytics landscape.

Furthering this analytic future solidly, Ingres last week announced the general availability of its VectorWise commercial analytic database that takes columnar technology even further. And, Ingres plans to provide an open source version before the year is out . Ingres describes VectorWise as next-generation analytic database technology as its announcement summarizes:

“Ingres VectorWise unlocks the power of modern commodity CPUs with a revolutionary database engine that leverages vector-based processing and on-chip memory to provide dramatic 10x - 70x performance gains over other databases.”

The performance claims are being validated by both customers and partners – and it just may be that Ingres has some groundbreaking new technology on its hands. With free downloads, trials, tools and documentation, Ingres is inviting the world to check it out. If it can use low-cost leadership along with super-fast processing to drive faster, interactive analytics and richer data visualization onto the laptops of more knowledge workers, I think Ingres VectorWise will have an important seat at the table of the future of analytics in the enterprise. Imagine that.

Brian Gentile
Chief Executive Officer

Tuesday, June 1, 2010

The Gulf Oil Spill should use an open source response

When BP CEO Tony Hayward first jumped into the spotlight in the days after the tragic explosion at the start of this massive leaking oil well catastrophe, he initially seemed reasoned, convicted and in control. As days passed and technique after technique failed to stem the flow of oil into the Gulf, the executive leadership and companies involved seem inept and even insufficiently skilled. At this point, there is rightly a great deal of blame being shared among the three companies involved and, indeed, within the Obama Administration. The net result of the spill will surely rank among the oil industry’s greatest environmental catastrophes.

So what could have been done differently during those first few days of the tragedy? What leadership methods might have been summoned to yield a different result? No surprise that I suggest an open source response.

What if on the third day after the explosion, Mr. Hayward called a news conference and explained the situation. It might have gone something like this...

“We currently estimate that more than 5,000 barrels of oil per day are leaking into the Gulf. Importantly, the source of the leak is, essentially, a broken pipe within our well, more than one mile below the surface of the ocean. At that depth there is almost no experience, even among the most tenured oil well and rig technicians, in capping and sealing a leak. We simply don’t know what we are up against. What we do have is extensive knowledge of the situation, tremendous and relevant experience capping leaking wells, and an enormous conviction to work quickly and tirelessly to prevent further environmental damage. So, as a team – BP, its affiliated partners and President Obama’s Administration – we’re calling on our global community to help. If you have ideas or expertise that might help us to more quickly implement a solution, we want to hear from you. No thought could be too big or too small. We’ll take ideas from any corner of the planet or any scientific discipline. And, we’ve set up a community web site to help receive and manage input. If you have ideas, I’m asking you to go to http://bp.savethegulf.org and let us know as soon as possible. Together, as a global community united to protect our environment, we can fix this tragedy more quickly and completely. Thank you.”

This leadership response has the benefit of authenticity (admitting vulnerability and limitation) and rallies support for a common cause rather than deflecting or trying to prove everything is all right (when it is clearly not). Most importantly, asking for global community involvement would almost assuredly yield ideas and possibilities to more quickly cap the flow of oil into this precious ecosystem. What could be more important?

Brian Gentile
Chief Executive Office