SAP In-memory – what is in the bag?
Five hours on an airplane on the way from SAPPHIRENOW back home in Silicon Valley should be enough to catch up with few levels of Crazy Birds on my Pre2, but as well with what have happened over the few days spent at the SAP main customers’ conference. As an SAP Mentor Initiative I was asked to participate in HANA Roundtables for SAP Premier Customer Network. It was an honor to co-lead these discussions together with the members of SAP Independent Executive Advisory Council (IEAC) and live legends in SAP world: Clive Weightman, who previously ran Global SAP Practice at Deloitte, ex-CIO of Colgate-Palmolive Ed Toben and ex-CIO of Valero Energy Hal Zesch.
These roundtable discussions showed that participants had huge interest in understanding the in-memory technology and its impact. Conversations like those show as well how wide the area is and how much is left to assumptions, partial information and misinterpretation. So, I am taking a challenge (another buzz-word from this SAPPHIRENOW) again to share with you my own collection of information and thoughts.
What is in the bag of terms?
First layer is in-memory technology – pretty wide term to describe everything that enables highly efficient storage and high performing processing of (sometimes substantial) data volumes. It will be multi-core CPUs, fast communication channels, right mixture of high-capacity types of storage, components’ redundancy for high-availability etc on the hardware side. On the software side it will be the whole range of data coding and compression techniques, sophisticated divide-and-conquer algorithms, cache aware programs, workload management mechanisms, fault recovery and so on. Hardware components can be commodity or proprietary; software techniques can be well-documented or patented. The point is: there are many vendors mixing and matching these components and techniques; SAP being one of those and calling its own set “SAP In-memory computing”. As explained in previous blogs, there are multiple products using SAP in-memory technology in one way or another.
Second layer is the platform: in discussed case it is SAP HANA™ delivered in a form of the appliance. “HANA” is being used pretty freely. Like in-memory based SAP BusinessOne running on Mac mini showed during Wednesday SAPPHIRENOW keynote was referenced as “HANA”, although Apple is not one of the six officially certified HANA vendor providers 😉
HANA is often confused as well for being the central software component, currently called SAP In-memory database (SAP IMDB). IMDB was previously referred as IMCE or ICE (you still can find these names in some documents, including my previous blogs). Actually SAP HANA 1.0 as a product comes with quite a few bundled software components: IMDB 1.0 Server, IMDB Clients (ODBC, JDBC, SQLDBC, ODBO), In-memory Computing Studio, Sybase Replication Server 15.5, Sybase Enterprise Connect Data Access, Sybase Replication Agent, HANA Load Controller 1.0, SAP Host Agent 7.20. Additionally the license includes SBO DataServices 4.0 with IMDB as the only target system. [Note: things are changing quickly and I am sharing the list for purpose of bundle illustration only. The actual configuration and licensing should always be checked with SAP.]
Now, HANA platform can be used to build the third layer – applications advanced by SAP In-memory computing: customer’s home-grown solutions or off-the-shelf products delivered by SAP or SAP partners. Products can be brand new, see Strategic Workforce Planning from SAP BusiunessObjects EPM portfolio, or existing, but re-worked to benefit from in-memory technology, like announced integration of SAP NetWeaver BW on top of HANA as the database and infrastructure.
PS. “AS-IS” applications will have to be added to HANA glossary as well. Those who watched Hasso’s keynote know what I am talking about 😉
PPS. In next blog I’ll try to add a clarity to the roadmap. At least as much as I can do.
-Vitaliy, aka @Sygyzmundovych