Big Data, Big Time

Series data, warehouse acceleration, and 4GLs

I just got an e-mail from the editors of this magazine, saying “Stuart, you’re late again. Where’s your article?” I replied, “Dear Editors, sorry. I have been busy.” I needed a topic—fast—so when I found out that this issue of IBM Data Management focused on big data, I started to wonder what big really means. Last October at the IBM Information On Demand (IOD) conference, we heard from two speakers in the general session. The first was a Visa executive who shared that the Visa credit card system handles the transactions of 1.8 billion credit cards—and they were not even all his wife’s.

Even if each credit card is used only once a month—and I tried valiantly last month to up the average, using mine 27 times—that’s 1.8 billion transactions a month.

Another speaker discussed the challenges of managing time series data, meaning that you track incoming data according to the time interval when it was recorded. At IOD, an electrical company representative said his company plans to use Informix to collect electrical usage readings every minute for each house. Each meter will produce at least 1,440 readings per day, 30 days per month, 12 times per year. That’s more than half a million readings per customer, per year. And if you took readings every second, that’s about 31 million per customer each year! Is that big data? Sounds kinda big to me.

What do these companies have in common? They’re IBM® Informix® database customers. So if a big data infrastructure must cope with TimeSeries, Informix is more than up to the job.

And if you need to accelerate the flow of all that information, the recently announced Informix Ultimate Warehouse Edition (IUWE) can help in warehouse and mixed workload environments. Based on Informix 11.7, IUWE uses a powerful columns-based engine that accelerates warehouse queries a hundredfold.

The best part is that the accelerator is completely transparent to the application, providing fast-response, ad hoc query processing without relying on I/O or partitioning. It compresses warehouse tables with an efficient algorithm and stores them in memory. The accelerator then accesses these tables without uncompressing data, so you don’t have index overhead. IUWE is also economical—for other people maxed out on their Visa cards—because it uses commodity hardware, requires no change to applications, and needs little database tuning.

And, here’s an announcement near to my heart because I started my career as an Informix 4GL programmer. IBM is expected to soon resell Genero from Four Js Software. Think of Genero as Informix4GL on steroids—the same language with a lot more graphical capabilities, including Windows or Web-based applications. Genero can even develop applications that run on an iPad or iPhone. Imagine running an old green-screen Informix 4GL in the palm of your hands. There’s now “an app for that.”

In addition, I recently asked readers why they use Informix. Fewer than 20 people answered, and I have editors to feed. Please go to and fill in the form. For just a few minutes of your day, you can help a columnist get on the straight-and-narrow with his editors, who are a mean and surly lot. Well, maybe not really mean and surly—just a little grumpy when I’m running late.