Vertica was one of the first so-called “analytical” systems (or columns with a massive-parallel computing architecture) of data warehouses, significantly speeding up not only data analysis, but also providing a number of other important advantages over string.
The architecture of traditional data warehouses was developed on the basis of line relational databases designed for real-time transaction processing (OLTP). Unlike OLTP tasks, analytics queries contain only a number of attributes, which means that for the bulk of analytics, the column architecture provides maximum performance and efficiency.
The secret of a column database is in its unsurpassed speed. In traditional relational databases, information is stored in rows, so even when your query requires data from a single column, the database scans the contents of all rows – each column in each row.
The second main advantage of column storage is that its architecture allows to reduce the amount of I / O operations – and this is the most important parameter for analytical processing.
The third advantage is efficient data compression, which provides 4-5 times higher performance than traditional databases.
Column storage is, of course, not a panacea, because there is no technology that would solve all the existing problems. But the disadvantages of column storage are transformed into advantages when it comes not to the transaction load, typical of accounting systems, and the analytical load. One of the features of the column platform, in particular, is that it stores each column in a separate set of files.
Deploying a column database in its basic configuration can slow down the addition and update of data, as well as delay – or complicate – their loading. To avoid such problems, columnar databases, in particular Vertica, implement optimized massive-parallel loading technologies and use relational approaches OLAP (ROLAP) or complex OLAP methods (MOLAP).
If the column architecture is optimal for data analysis, then the massive-parallel architecture is for scaling analytical processing.
Want to know more? If you have any questions, please contact email@example.com and write in the comments what you are interested in about Vertica!