SAP already saw the qualities of in-memory some time ago. The company acquired Transact In Memory in 2005 which was already doing commercial memory-centric relational database solutions. Under SAP, the eventual technology developed was named HANA, short for "high-performance analytical appliance". It is both a revolution and evolution in database management as it utilises in-memory-capable hardware with the database management functions of SAP software.
This type of technology can be such a game-changer for companies who rely on data-intensive analysis especially if they have volumes of customer data that arrive every second. Attesting to this use of in-memory database are companies like Nongfu Spring in China and Nomura Research Institute from Japan, both of which have tried, tested, and succeeded in implementing in-memory database systems.
A Tale of Two Companies
Nongfu Spring is a leading distributor of water and fruit drinks in China, and is currently holding the largest share of the market.
The company faced many challenges such as, its sales team was collecting point-of-sale terminal data in different ways. The large volume of data meant it would take a day to run ETL (extract, transform and load data) and another day to receive these channels and POS (point-of-sale) data. That hampered Nongfu executives' ability to make informed decisions and take action, according to Patrick Hoo, CIO, Nongfu Spring.
Nongfu Spring successfully went live with SAP HANA last year, and upon implementation, they found that it met the three goals they had set out to achieve. The first goal was to accelerate its displaying reports. Nongfu Spring did the same report, running on the same script, in both HANA and Oracle. They found that the same script was 200 to 300 times faster on SAP HANA than on PL SQL on Oracle. This result was consistent over all 150 reports.
The second goal was to produce highly efficient business logic calculations. They transferred over 50 stored procedures and functions from Oracle's Data Mart to SAP HANA and compared the calculation speeds. Using a freight report example, what took 24 hours on Oracle only took 37 seconds after it was optimised on SAP HANA.
The third goal was to achieve real time data replication and data synchronisation. In the past, captured data was displayed 24 hours after ETL. But now, real time data is displayed instantly.
The SAP system, Hoo said, minimised the possibility of mistakes, as well as reduces Nongfu Spring's maintenance costs. It proved to be a comprehensive and sophisticated data solution that Nongfu Spring will continue rolling out. The company intends to introduce a mobile distribution system, leveraging HANA.
As one of Japan's leading consulting firms, Nomura Research Institute (NRI) is a driving force in the economy of the country. In finance, logistics, industry, and the public sector, this highly professional institute has developed a reputation for cutting-edge solutions and high-end services.
Sign up for Computerworld eNewsletters.