What is a typical reaction to high-performance analytics? Consider these scenarios:
- What do investment bankers say when you tell them the risk calculations they used to do every three days on a subset of their portfolio can now be done within minutes - all while analyzing the full portfolio?
- What do retailers say when you tell them they can optimize prices for all items in stock, customize prices per store and even predict pricing changes over the next few seasons?
- What do patients say when you tell them their entire medical history can be stored, analyzed and compared to similar family and geographic healthcare data to help predict and prevent disease?
- What do government executives say when you tell them they can use public record data to find connections in random activities and identify fraud rings that could save tax payers tens of millions of dollars?
They say, "That's incredible!" Just look at the numbers in these ads. They really are incredible:
There's a sense of incredulity that comes along with these types of promises, and that's understandable. When some of our high-performance analytics (HPA) experts presented these - and other - scenarios to customers recently at a SAS customer advisory board meeting, the initial reaction was skeptical. But after the developers discussed the technology at a deeper level, even the skeptics started to understand - and then they got excited about how they might use the technology in their organizations.
Understanding what you need: analytics and speed
Solving the problems above require two things: advanced analytics and fast computing capabilities. Just getting the data quickly won't give you those incredible results. And analytics alone won't give you those results either. You need high-performance analytics.
In other words, each of the examples above require some type of advanced analytic technique: optimization, forecasting, text anaytics or social network analysis. But they also require implementing those techniques at faster speeds, on larger amounts of data and on data collections that are changing by the minute.
When you move the analytics to work inside or alongside the database, you overcome all of those issues. That's what high-performance analytics does.
In a recent National Post interview with Jim Goodnight, the SAS CEO explains it like this:
There's a lot of business processes that will be changing because of the speed at which we can do analytics; using a thousand processes in parallel to do these computations can make it possible to do huge problems that we would never have been able to do before because it would take too long on a single processor.
Did you catch that? One thousand processes in parallel. That's incredible! But he's serious. Goodnight has been testing these calculations himself on his own configuration of blade servers.
Reduce your large analytic problems to a series of smaller problems
How does it actually work? In "The promise of high-performance analytics," Paul Kent, VP of Platform R&D SAS simplified the answer:
We’re not inviting data to come to us so we can munch on it anymore. We’re finding clever ways to go where the data are, and move the work out to all the different slices of data as it exists.
Paul went on to explain how multiple calculations are conducted on different nodes simultaneously and brought back together for a final answer. That's a big part of how HPA gets its speed: it breaks larger problems down into smaller pieces.
So, really, it's not that hard to believe when you think about all the large tasks you accomplish in a similar manner all the time. We solve big problems at work and at home all the time. We just break them down, spread the work around and come together in the end to create something amazing. Yes, that's incredible too - but it's also very believable.
What incredible things can your organization accomplish if you could break some of your largest problems down into manageable chunks to be solved simultaneously? We'll do the chunking and the analyzying. Just bring us the problems. Or, as Rhadika Kulkarni said at a recent SAS conference, "Bring us your challenging situations. Bring us the problems you thought were not solvable. That's what we revel in."