I have been working on SAP HANA for long time. First call with SAP’s R&D was in 2010, when HANA was yet to be named and they were still calling it Hasso DB (for Hasso Plattner, co-founder of SAP). Now I come to think of it, even if HANA officially means High performance ANalytical Application, it is said that the real meaning is HAsso New Architecture, who knows…
My first official SAP meeting on HANA was in Walldorf (German SAP headquarters) in March 2011. There were 200 of us from IS and HW vendors who went to discover HANA and hear the marketing and some technical speeches.
During this event, SAP heavily emphasized the fact that computers and memory chips are cheap, that they can address 64 bits, that columnar compression allows storing lot of data in RAN, that HANA was very fast …
Put simply, SAP HANA is a SQL database where all data is stored in-memory. Of course, it is much more complex:
- HANA is sold as an appliance with dedicated hardware (provided by IBM, HP, Fujitsu, Cisco, Dell and more recently Hitachi)
- It contains Solid State Disks (SSD) to store the updating logs (to respect ACID properties)
- All data is regularly saved to classic Hard Disk Drives (HDD) to allow a reboot in case of a power cut for instance
- The HANA license includes a modeling tool
- The HANA license includes a loading or a replication tool.
SAP’s description of SAP HANA has partially evolved. They still put emphasis on these technical facts, but they have added a very interesting and new fact; HANA is not only a fast database, it is part of a whole solution that improves business. Moreover, it means SAP doesn’t target customers or prospects with the speed of HANA; they come with a solution (e.g. budgeting, profit computing, planning, human resource management, BW improvement…) with new capabilities thanks to the use of the HANA database under that.
I like this new description, as for once, the way to justify (and to sell) a database is linked to business value, instead of focusing on technical things. I saw that in the past only for multidimensional databases like ESSbase. This is not surprising as one of my previous posts explains in-memory databases are killing OLAP databases.
But in my mind, the real value of HANA is still somewhere else. The value of HANA is that we are able to do a complex calculation very fast, fast enough to be part of the transactional business process.
A “cold BI process” (i.e. done out of transactional process, like a dedicated data warehouse or an EPM solution) works fine with the classic very large database technologies (like Teradata, Oracle Exadata, HP Vertica, EMC Greenplum, SAP Sybase IQ…), even if having it on an in-memory database may improve it.
But HANA makes it possible to have BI technology not only in real-time, but as part of any business process, like allowing a Telco or a banking company to do fraud analytics for each call or transfer, like allowing a manufacturer to reorder immediately its supply chain for each new order or even quote, like allowing a retailer to manage the inventory depending on Point of Sales activity in real-time. This is what HANA will provide. Are you ready to redesign your business processes? Are you ready to improve them? Are you ready to understand the impacts HANA may have on your business? Are you ready to manage the changes? HANA is not just a technology change; it is a business breakthrough technology.
Maybe you’re wondering why I’ve put a waterfall image; because Hana is also a very nice place in Hawaii.