Dave's Adventures in Business Intelligence

A few weeks ago I attended the quarterly meeting for the Dallas/Fort Worth ASUG chapter. I didn’t get to stay for the entire day, but I did get to hear the keynote by Dr. Jeffrey Word about HANA. The talk was less about the technical aspects of HANA and more about the genesis of the idea. He started with a very interesting comparison. It seems that HANA is SAP’s iPod.

iPod Evolution

The original iPod had a small hard drive (5GB) to store music files. However, even as small as it was, the hard drive (and the battery to keep it running) took most of the interior space of the case, leaving very little room for interface elements like buttons or a screen. As a result, the Apple engineers came up with the click wheel, an interface that provided all of the functions required to browse and play music. Later generations of the iPod were able to take advantage of better (and smaller) technologies and provide a color screen, larger hard drives (80GB), better battery life, and so on. But the framework was still restricted by the hard drive component. The color screen was still small, and the click wheel was still used to browse the library. Most functions were very easy, but searching was still painful due to the lack of a keyboard.

In 2007 the iPod became available in an 8GB “touch” model. The iPod Touch was very similar to the iPhone. It had a full-sized touch screen interface, meaning no more click wheel. How was this possible? The guts of the iPod Touch don’t contain a hard-drive. Instead the Touch relies on non-volatile memory for storage. Memory requires less physical space which allows for more flexibility in the overall system design. Memory requires less power which allows the battery to be smaller but still provide lengthy life. Memory does not suffer from physical abuse or accidents such as drops, making the device more reliable. And with the full-sized touch screen, the iPod Touch could do more than play music, it opened up an entirely new world of games and other applications. The trade off? Only 8GB of space. My iPod Classic has 160GB of disk space for music and other data files.

Today the iPod Touch can be ordered with up to 64GB of storage. It includes a high resolution camera, can take high-definition video, and runs thousands upon thousands of games and other applications from the Apple store, very few of which were written by Apple engineers.

What does any of this have to do with SAP, and HANA specifically?

Memory Removes Restrictions

The iPod Touch (and its cousin the iPhone) would not exist without the in-memory storage to replace the bulky and battery-draining hard drive. By removing the hard drive, Apple opened the iPod up to all sorts of possibilities. In the beginning there weren’t many applications; Apple provided only a few to start with, and there was no developer kit for outside programmers to use. As Apple grew more comfortable with the framework they started to open up the devices for outsiders to play with. I’m sure they had no idea that Angry Birds was on the horizon. 😆

HANA today for SAP is what the first generation iPod Touch was for Apple. There aren’t a lot of applications written specifically to take advantage of the architecture just yet, and SAP has not provided any sort of SDK. (HANA supports both MDX and SQL which means most standard database calls should already work.) But SAP is already working on porting their planning applications over to HANA, and more are sure to come. One of the customer testimonials at SAPPHIRE was from a company that apparently used the initial versions of the in-memory planning application on a test system, and they were blown away by the responsiveness of the system.

HANA tomorrow … is a 64GB iPod Touch? Memory capacity continues to increase. Processor power continues to increase. As HANA matures the expectation is that more and more applications will be written to take advantage of the speed and power of the in-memory architecture. Some of those apps will certainly be to support the ERP systems, but certainly some will be in the business intelligence space as well. One of the more lofty projections suggests that the entire concept of extracting data to a warehouse for reporting will become obsolete. Data warehouses came about because we needed to be able to run massive number-crunching reports without impacting the day-to-day operations of our base systems. If those base systems are running on HANA, the theory is there will be no impact. We’ll run reports on “live” data, at the same time we’re running our transactional systems. Think of how much simpler (and less costly) this could be. No more ETL. No more warehouse hardware. No more reporting off of day-old data. The entire IT infrastructure can collapse down to operational systems.

Wow.

We’re certainly a long way from tossing all of our warehouses aside and reporting off of real-time data. Apple took about ten years to get from the original iPod to the iPod Touch 64GB we see today; the first touch appeared roughly in the middle of that evolution. HANA has roots in TREX, a text search engine that SAP built back in the late 90’s. The first incarnation of the in-memory database came in 2005 with the release of the accelerated version of BW. That engine was a specific-purpose implementation of the in-memory engine. Today HANA is a “generic” version of the in-memory systems used to accelerate BW, suitable for plugging in behind the scenes anywhere a database is required.

But someday, a couple of people in a garage just might come up with the Angry Birds of HANA. That would be something to see. 😎