The world of databases has fundamentally remained the same since the 70s while every other layer of the stack from chips to UI has seen dramatic improvements in price and performance. However, in the last 2 years, with the advent of HANA and HADOOP, we see a radical rethink of this layer fundamentally challenging size, variety, speed and simplicity all simultaneously. A large part of the credit for these advances should be given to the constant acceleration of innovation in the world of hardware but also to new software approaches that took advantage of them. What has also changed is the interconnectedness of the global businesses that make use of these technologies. Today, businesses need to take action within the window of opportunity, not only capture transactions but bring the collective power of data to inform every interaction in real-time across business networks.
Incremental vs. Radical Rethinking of Databases
We recognized the importance of data processing speed several years ago with in-memory based add-ons to disk based databases such as Live Cache and Business Warehouse Accelerator. However, improving speed alone was not sufficient – you have to simultaneously innovate on multiple dimensions – go across broad data types including machine and multi-structured data, do deep analyses such as predictive algorithms, work on fresh data without latency and at interactive speeds that is as fast as the human mind while simplifying IT landscapes from redundant stores, indexes and aggregates.
Recently IBM announced BLU accelerator for DB2 which does query acceleration of manually selected tables in DB2 in batch mode using in-memory and columnar techniques. Previously Oracle announced Exalytics, Microsoft announced SQL Server 14 (Hekaton), Teradata announced Teradata Intelligent Memory. It is quite a welcome sign that the database industry has recognized the new category of in-memory data platforms. However, all these attempts fall far short as they all have a leg in the outdated disk based architecture for transactions/datawarehouse and are trying to bolt on in-memory capabilities for faster analytics. Most of them don’t capture transactions in memory and don’t do analysis on live transactions in real time with zero latency and simplification of landscape. Incremental approaches are simply missing the point.
To be relevant in today’s world – existing database solutions have to be reimagined completely. Ford did not come up with a faster horse carriage; he reinvented transportation with the car. And that is why we invested so heavily in SAP HANA and started with a fundamentally different design center that changes how we capture, store and process data. Why should customers accept separation of data processing workloads such as OLTP and OLAP as the norm? Why should they be forced to spend on domain-specific databases for predictive, natural language or spatial data processing, especially when all these capabilities are required by the same application? Why can’t we extend the database to include application platform capabilities to push down data-intensive tasks such as planning, rules, and calculations within the database? Why can’t we build all these capabilities on a single platform?
A Re-imagined Platform
SAP HANA is not just a database; it is a completely re-imagined platform for intelligent real-time applications. It provides customers with converged capabilities to transact AND analyze fresh data of any kind, eliminates batch processing, streamlines multiple data and application logic processing workloads (without requiring multiple data copies), eliminates redundant layers, scales to more than 1 PB in main memory and still has rapid sub-second responses.
We witnessed a similar transformation with the smart phone platform. When you converge multiple capabilities such as GPS, internet access, music, voice, camera capabilities into a single phone device, amazing applications are created by extraordinary people across the world.
In addition, HANA is a full-fledged platform with natural language processing, a rules engine, a planning engine, a native application server and a native web server built in (and SAP doesn’t charge extra add-on licensing in order to use this functionality). SAP HANA also comes with libraries of embedded business, statistical, and predictive algorithms and geo-spatial capabilities. SAP HANA erases the line between OLTP and OLAP and provides enterprise applications the mixed workload environment they now need for real-time information processing across transaction and analytics, using structured and unstructured data, bringing context and prescriptive recommendations to users the instant they need it. I have no doubts in my mind that this is the platform for applications of the future.
SAP HANA is now already at 1500+ customers and is used to solve some of the most sophisticated problems that customers face, not only for SAP applications and analytics, but more importantly for new applications such as those from 400+ startups, genomics and personalized marketing.
Our customers successfully use SAP HANA to
- Make real time currency hedging calculations for big banks
- Do complex Reverse BOM at machine manufacturers
- Deliver mobile applications that make real time recommendations for retail customers while they are shopping based on real-time geo location
- Combine ERP information with machine data for predictive maintenance
- Help scouts identify the best sports talent by deep analysis of multiple data sets
- Run real time MRP to reduce working capital requirements and so on.
- …and of-course run the world’s most demanding applications SAP Business Suite and BW
All this is done by SAP HANA in real time on a converged platform. None of these high value solutions can be done by the incremental approach of using selective query acceleration with an in-memory add on followed by the traditional vendors. SAP has nothing but great respect for IBM, Oracle, Teradata and others, but it’s time to lead the industry in the direction of innovation to better serve our customers.