SAP Person: “HANA can run OLTP and OLAP from one copy of the data, speeding things up by thousands of times, dispensing with the use of aggregated data and indexes, shrink system sizes and costs through landscape simplification by 10x or more. It allows you to radically simplify what you do, making users and developers 2-3x more productive, becoming 6x more agile or even more. It allows you to flex your organisation in response to threats or opportunities much more easily than you’ve thought possible … “
Customer: “Whoa! I’ve been in IT for 20 years and frankly those things are impossible, I don’t believe it, they’re simply too good to be true, you are either pulling my leg, or just plain crazy”.
My friend Thijs in the Netherlands, who studied psychology, informed me that this is a well-understood phenomenon. We resist information that contradicts our existing mental models, and the more challenging information we are presented with, the more it invokes resistance.
What is needed to combat this problem is patient consideration on why some of our fundamental assumptions learnt in the past, over many years, may actually no longer be true. To do this for HANA, and to understand why HANA is so different, takes a bit of time, about 90 minutes to be precise. But if we take sufficient time it allows us to go behind the startling claims and thoughtfully step through exactly where HANA came from, what its original aims were, and how it was developed. If we do this then these claims don’t need to be explained, but they become obvious. I found myself doing this over and over again, it turned into a presentation, and I was then asked to turn it into the video which can be found here:
Link to “HANA The Why” YouTube Video
You’ll need 98 minutes (and probably some coffee) to view it, it explains how:
- Ten years ago SAP needed to overcome complexity and cost to find a way to pull ahead and stay ahead of the market
- The strategic answer was to simplify everything we do, and in so becoming more agile and productive, that way we would be able to out-innovate our competition
- Hasso Plattner, our chairman, kicked-off a project of pure research (having founded his own university) to see if we could invent the enterprise system of the future
- The researchers succeeded. They found a simpler and more elegant way to build enterprise systems
- Over twenty major techniques, new and modified, were needed. Interestingly only one was about using memory rather than disks, there was much more to it than that
- Exploiting modern microchip features, that appeared from about 2005 onward, was key – this could yield speed improvements of 100,000 times(sic) or more
- The dramatic speed increase eliminated the need for aggregates and indexes in applications, other research told us these caused 60-95% or more of application complexity – a massive simplification. Speedup is so great that even after dynamically producing aggregates on the fly there is still plenty of performance spare to be hundreds or thousands of times faster than traditional systems
- But this had to be based on column stores, notoriously difficult to update, so they invented a unique technique – suitable for running thousands or millions of transactions per second
- Thus Hybrid Transactional and Analytical Processing (HTAP) became possible, with both types of processing being done on the same single copy of the data, which allows for huge landscape simplification. Operational reporting can now return to the operational systems with large systems savings and the agility of reporting on real time operational data.
- The speedup also benefited text processing, spatial processing, planning, predictive, and graph processing, all could be mixed together at will, in-memory, and in-parallel. This allowed for greater agility and productivity, not just through the speed of execution but by having all these techniques at your fingertips ready to be used and combined.
- The SAP Development organisation, working 24/7 ‘following the sun’ in multiple centres around the world expanded the research out into today’s fully featured HANA Platform
- This work produced an enterprise system that is fundamentally simpler, and that enables high productivity, much greater agility, higher performance and lower TCO.
There are of course many more details on the features and functions of HANA, and there are plenty of SAP, Partner and Customer materials available to explain them, I don’t reproduce these in the video, which focuses on answering the questions “Why was HANA invented, Why is it different?” I think you’ll find knowing this makes it much easier to slot in the details later.
For more detail on specific aspects of HANA you’ll find lots of material on the internet, and in particular the video tutorials of the SAP HANA Academy which can be found here
Link to the SAP HANA Academy
Acknowledgements: As you’d expect, as this evolved, I gathered information from many sources, mainly the talks, books and academic papers of Hasso Plattner which are both inspirational and beautifully reasoned. Also the many discussions with colleagues such as Mark Brown who collaborated on the first fifty or so iterations, Dan Holle, Andre Borchert, Thijs Elling, Alexis Fouquier, Carl Streatfield, Paul Barker, Erich Schneider, Juergen Butsmann, Ian Henry, Guneet Wadwha, Mark Mitchell, Dimitris Kouvaras, Erich Schneider, Lothar Henkes, Daniel Rutschmann, Irfan Khan, Chris Hallenbeck, Scott Sleeper, Walt Muntzenberger, Tim Hardy, Paul Everett, Andrew de Rozairo, Simon Harrison, James Willis, Michael Longden, and many others, including SAP Customers and Partners, many thanks to them for their input, wittingly or otherwise, and the many stimulating discussions around this topic.
Thanks especially to Mohamed Ahmed and Rachael Coates in the UK for their patience and perseverance with the video editing and Ben Benitez for help with the blog.
Lastly, thanks to Simon Sinek who’s excellent book “Start with Why”, got me thinking about explaining SAP HANA this way. Link to: YouTube video Simon Sinek “Start With Why”
As ever, any errors are mine. I hope you find this useful.
Henry Cook, March 2016
VN:F [1.9.22_1171]HANA The Why,