Sap Hana

Based on the response of question 3, SAP HANA has the following advantages over the other competitors:

• SAP HANA takes full advantage of the new hardware technologies by combining columnar data storage, massively parallel processing and in memory computing by using optimized software design.
• SAP HANA performs application logic using predictive, natural language and spatial processing. It comes with built-in libraries of embedded business, statistical, and predictive algorithms. Therefore it simplifies the overall architecture, increases processing speed and the development of application.
• SAP HANA bridges the gap between OLTP and OLAP and provides real-time information processing across transaction and analytics.

Promising Features:

Row and Column Store

SAP (2013) describes the two data processing approaches for SAP HANA:

• The column-based approach stores relational data in columns and is optimized for holding large volumes of data, which are aggregated and used in analytical operations.
• The row-based approach, stores relational data in rows. This row based approach is optimized for write operations but it provides a lower compression rate than the column based approach, and its query performance is significantly decreased compared to the column-based store.

SAP HANA allows selecting per-table basis at the time of creation of table to store data.  Tables that are in the row-store are loaded into memory at start-up time, whereas tables in the column-store can either be loaded at start-up or on demand, during normal operation of the database.SAP(2013)

Compression

SAP (2013) describes SAP HANA’s compressive feature as a set that allows simultaneous handling of real- time transactions and analytical workloads with extreme speed.

The columnar approach of storage provides higher efficiency in compression of data which reduces the cost of data to be stored in memory by the SAP HANA database. This results in high processing speed required to perform searching operations and also executing calculations.

SAP (2012) conducted SAP HANA’s Performance tests. It states that data compression occurs during the data loading process which demonstrates compression rate greater than 20 times than traditional disk based systems; the 100 TB SD data set was reduced to a trim 3.78 TB.

Data in the column table can have a twofold compression:

• Dictionary Compression: This is the default method of compression that is applicable to all the columns. It is achieved by mapping all the distinct column values to consecutive numbers, by which instead of the actual value being stored, consecutive numbers are stored, as they occupy much smaller memory space compared to the actual values.

• Advanced Compression: Apart from the default compression of dictionary compression method, each column can be further compressed by using various compression methods such as prefix encoding, , sparse encoding, cluster encoding and indirect encoding. SAP HANA decides the appropriate compression methods for the column with the help of compression algorithms.

Scalability 

SAP HANA supports scalability in two dimensions. They are scale-up (increase the size of hardware in a database server) and scale-out (increase the number of database servers on a single database system). Although SAP HANA supports both the versions of scalability the Scale-up option is limited due to the hardware that is available. SAP HANA supports scale-out approach, by using a mechanism that will distribute the execution of the query upon the main memories of the multiple database servers in the same database system. SAP HANA ensures that each database server can work independently on its own set of data by reducing the data transfer between the servers at the same time (Bendelac 2012).