Application And Reasearch Based On Sap Hana In-memory Computing Database Technology

These years, more and more companies have requirements to compute and analyze bigdata. However, if they still rely on current ERP or BI system, it’d be frustrating to deal withthe job for computing and analyzing the huge amount of data. The reason is that, currentlyalmost DBMS is based on traditional disk array. If we need OLAP to perform some logicaloperation with mass data, for example, like aggregation operation, it will cause a too longoperating period for analytical reporting. Even if we got the data we want, we have beendisappointed by losing the real time data. Facing the fact, we start to pay high attention to theIn-Memory Database Technology. Then, in terms of the current developing status ofIn-Memory Database, how about its utilization? Is it worth commercial popularization? Whatwe talking here is trying to launch the discussion about the technical application research.In-Memory Computing, or say, In-Memory Database, is not new to us. In certain aspects,In-Memory Computing has a big advantage to the database based on disk array. This articlediscussed a series of core technologies of In-Memory Computing, which fostered thedevelopment of In-Memory Computing Database: By the high-speed development ofhardware, the mass application of parallel control and multi-processors technology becomemuch more perfect. Although the cost of main memory is going down, comparing with disk, itstill pretty expensive. So the technology of data compression is also very important. Thecolumn compression helps to increase the information density. These data compressiontechnology improves the system performance. On the other hand, it also keeps the memoryconsumption in a low level. SAP HANA is a typical product of in-memory database, which consists of row store, column store and mix store technology by comparing with traditionaldisk array database. Column store ignores those unnecessary columns, which loads much lessdata rather than the traditional row store.Long term experience makes the author believe that the success of a new technologydepends on business connectivity rather than only replying on the advanced technology. Theborn of SAP HANA provides a new chance and also makes it possible to compare with thetraditional disk array based DBMS. As a new tool, HANA also need to be improved andenhanced by widely practices. Through the practice in3HANA projects, the authordemonstrated how to do the modeling according to real business scenario and implementefficiency quarterly sales analysis report; secondly, it represents how to utilize the HANAinside-based Application Function Library (AFL) by SAP HANA Studio based on the realbusiness scenario and then realize lite sales forecast analysis, which proves the feasibility offast business and technic deployment; finally, by a design of data compression and dataloading experiment, it’s tested and analyzed in the performance of compressing and loadingthe data with variant dispersion. It is concluded that In-Memory Computing Database, such asSAP HANA, has not only technical advantage but also disadvantage. But it’s truly promisingand portending a bright landscape in the next decade.Based on all the above, In-Memory Database is a direction of development, but it stillfaces some challenges and bottlenecks. Such kinds of problems as, the high investment in theearly stage; programmers have to change the methodology which is used in traditional diskarray database; the new application in business intelligence and emerging field of cloudcomputing. All the problems are all need to be discovered, research and develop in further.