With the delivery of the latest service pack for the SAP HANA in-memory computing platform, SAP has begun putting the tools in place that developers need to build applications on top of the HANA API.
Service pack 7 of SAP HANA includes a Web-based integrated development environment that SAP expects will become a critical component of its campaign to recruit developers to the platform. SAP claims to already have more than 1,000 startup companies building applications on top of HANA.
HANA is based on a columnar database that SAP HANA can use to run both transactional and analytics applications. That capability dramatically reduces the amount of IT infrastructure needed to run transactional and analytics applications that previously ran on separate platforms.
Vishal Sikka, a member of the SAP executive board that also oversees technology and innovation for SAP, says the power of HANA platform comes first from the fact that it doesn’t matter whether it's operational or analytic data. The end result, says Sikka, is “computer revolution" where applications run orders of magnitude faster than platforms that rely on magnetic disks for primary storage. The second reason is that SAP HANA compresses data in a way that dramatically reduces the amount of infrastructure needed to support any given application. Competitors that focus only the in-memory computing part of that paradigm, says Sikka, are missing the point of the extreme simplification enabled by HANA.
For developers, the implications of converging transactional and analytics workloads within the same application could be profound. For the first time, analytic applications will be informed of changes to transaction processing patterns in real time, which in turn will allow them to change attributes of the business process in prescriptive manner, also in real time. The end result is a cycle of constant tuning of a business process.
Better still, because HANA allows each application to work against raw data stored in the columnar database, Sikka says there is no need to duplicate data across multiple applications. That winds up eliminating 30 to 40 percent of the data stored across distinct applications, and shrinking the size of the database by almost 50 percent, he says.
Right now HANA can process 3.5 billion scans per second per core using the latest generation of Intel Xeon processors. SAP expects to be able to achieve 5 billion scans per second per core across 16 cores per processor. Further on, Sikka says SAP is also working with Intel on next generation non-volatile memory technologies to boost that performance in the future even higher.
The success of a platform such as HANA ultimately rides on developer acceptance, which is one of the primary reasons that SAP is building out an API management platform. While the platform itself has been around for a while, SAP only now is putting the tools in the hands of developers that will ultimately determine whether HANA acceptance will expand beyond the applications that SAP itself develops for it.