Bringing the world of computing up to speed
Vishal Sikka: "We are working on a strategy that makes us an enterprise platform company, with Hana as the core of that platform." PHOTOGRAPH: ERIC LUKE
When it comes to career achievements, chief technology officer of SAP Vishal Sikka ranks the creation of his “baby girl” Hana as among the top so far.
Hana is 19 months old, he says, and has the potential to be a disruptive influence on the technology industry. In fact, the application of Hana has wider repercussions for the world as a whole.
“Hana itself is a rethinking of the database,” says Sikka, adding that it could be responsible for a “massive simplification” of the IT landscape.
The new technology takes advantage of a number of different developments in the IT landscape, primarily the low cost of main memory (RAM) and the power of multicore processors.
It’s probably not before time; the relational database has been around for about 25 years and, as Sikka points out, the underlying nature of hardware was significantly different than it is now.
“For instance, the benefits of Moore’s Law were being achieved primarily through increasing the performance of CPUs, which were at the time single core CPUs,” he explains.
“Main memory was vastly more expensive and smaller than it is now. So around 2003 when we started thinking about this, our assumption was that the next generation database has to be natively and massively parallel, and has to take maximum advantage of main memory.
“Hana optimises around those principles, that all data can be in memory now. And we can take advantage of the massive amount of CPU power that we have available. The end result of that is extreme high performance in database processing, data processing, information processing.”
The basic premise of Hana is that it helps to speed up calculations. On the face of it, that doesn’t look like much, but when it’s applied to real world situations, the potential is far-reaching.
For example, the processing of medical tests could be vastly accelerated, cut from weeks at present to almost real time.
“There is no reason for this to take seven to eight weeks or 30 days. Thirty days for someone who is diagnosed with cancer is a very long time,” says Sikka. “We think there is no reason for this to take 30 days, that it should happen while the patient is sitting with the doctor.”
SAP is breaking out of its traditional mold. The company is working with researchers at Stanford University – where Sikka completed his doctoral thesis – to rethink genome analysis.
“We do the alignments eight and a half times faster than the fastest technique today,” he says. “And we do annotations and we do the analysis of annotations up to 6,000 times faster. So that means we can shrink the time it takes to analyse a person’s genome dramatically, and that is something that is very exciting but also very purposeful.”
One firm is using Hana to monitor healthcare data in elderly patients – diabetes, blood pressure, heart rate – with both automatic monitoring and proactive reporting of this data.
Another firm in France is using Hana to run calculations on stars, replacing work that would previously have had to be carried out on a Nasa supercomputer.
“We have data on hundreds of thousands of stars now, and we make 3D models in computers of the sky,” says Sikka. “Later this year there will be the satellite that will go up and bring back data on a billion stars.
“When you are looking at a 3D map of a billion-plus stars in the universe, the angle at which the human eye sees that 3D field, the eye can absorb maybe 100 to 150 stars. So which 150 stars to paint for you and to what degree of luminosity, how bright they are and so on, these calculations are done on Hana.