SAP HANA, standing for high-performance analytical application, is an in-memory, column-oriented, relational database management system created and marketed by SAP SE. As a database server, its primary function is to store and retrieve data on requests by the applications. It also performs advanced analytics (predictive analytics, text analytics, text search, spatial data processing, streaming analytics, graph data processing) and includes Extract-Transform-Load (ETL) capabilities plus an application server.
The key difference between HANA and earlier SAP systems is that it is a column-oriented, in-memory database combining OLAP and OLTP actions and resulting as an OLTAP system. The advantage of using main memory instead of disks is faster data access, and high-speed querying and processing. But it is a costlier form of data storage. Observing data access patterns, at most 85% of data in an enterprise system may be seldom accessed. Therefore it can be economical to store frequently accessed data in-memory while the less used data is kept on disk, an approach SAP have named as “Dynamic tiering”.
Column-oriented systems store a column’s entire data in the same location, rather than storing all data within a single row in the same location. This can improve performance for OLAP queries on large datasets, and allow higher vertical compression of similar types of data in one column. If the read times for column-store data is fast enough, combined views of the data can be formed quickly, removing the need for keeping aggregate views and the associated data redundancy.
Although OLTP makes use of the traditional row-oriented systems, both OLAP and OLTP can make use of in-memory storage, to develop hybrid systems thereby removing the need to manage separate systems for the two different kinds of operations.
The index server in SAP HANA architecture performs session management, transaction management, authorization, and command processing. The database has a row store and a columnar store. Users can generate tables using either store, but the columnar one has added capabilities and is preferred. The index server also achieves persistence between cached memory images of database objects, permanent storage files, and log files.
SAP HANA Information Modelling/SAP HANA Data Modelling also constitutes HANA application development. Modelling is the mode to expose operational data to the end-user. Reusable virtual objects (called calculation views) are employed in the modelling process.
Another important point worth mentioning is that SAP HANA is Geospatial DBMS. It can store, process and visualize geometrical and spatial data as well as allow operations like distance calculations and determination of union and intersection of objects. Moreover, you can integrate geo-data with any existing structured and unstructured data.
Coming to its analytical abilities, Stechies.com states that SAP HANA carries a number of analytic engines for several kinds of data processing. Its library called, the Business Function Library (BFL) includes a number of algorithms made available to attend to common business data processing algorithms like asset depreciation, moving average and rolling forecast. Another library within SAP HANA known as the Predictive Analytics Library consists of native algorithms for calculating common statistical measures in the areas of clustering, classification and time series analysis.