Sadas Engine is a relational DBMS specifically designed to optimize the execution times of analytical operations on huge volumes of static data, like those in a Data Warehouse.

The most relevant difference is the physical organization of data: traditional databases store data by rows, whereas Sadas Engine stores data by columns.

The traditional row-based approach is certainly the best in transactional environments, where it is usually necessary to access many attributes connected to a limited number of items.
In OLAP applications, instead, a limited number of attributes connected to very high volumes of data is usually analyzed; in this case, a column-based organization is much more efficient. With a columnar organization it is also possible to have a more efficient compression of data, thus further reducing I/O times of the queries.

The main factor is the column-based organization, but Sadas Engine also uses other techniques based on the static nature of the data managed (which are fundamentally read-only):
– the creation of specific indexes structures, based on proprietary algorithms, that accelerate the processing of big files with appropriate redundancies;
– an approach of automatic adaptation of the structures depending on the most frequently run queries (learn by usage). Basically, the execution plans deriving from queries are continuously analyzed and, for the most frequent ones, additional specific indexes and data structures are created automatically, thus speeding up their processing.

After a certain utilization period, the most common queries will be executed even more quickly than they were immediately after the installation. It is however remarkable the fact that the original execution time with Sadas Engine is considerably quicker than the execution time with traditional DBMSs.

To answer this question it was used the TPC Benchmark™H (TPC-H), a standard for the evaluation of DWH systems. The performances obtained were from 5 to 75 times better, depending on the type of query that had been run. These values refer to the “initial” processing and do not consider the further improvements that can be obtained progressively thanks to the “learn by usage” techniques.

There are two main kinds of benefits:
– operational: users can obtain the results they need much more quickly, and this sometimes has an impact on the actual possibility of doing some particular analyses.
– economic: thanks to the greater operational efficiency of Sadas Engine, the hardware and software infrastructures necessary for the DWH environment are much cheaper than those necessary with traditional tools.

Sadas Engine was specifically designed to manage big volumes of data; these are in fact the cases in which traditional tools sometimes have problems with performances, and in which it is logical to introduce alternative technologies.