Differential buffer for a relational column store in-memory database

At the present the financial and the analytical reporting has taken importance over the operational reporting. The main difference is that operational reporting focus on day-to-day operations and requires data on the detail of transactions, while the financial and analytical reporting focus on long term operations and uses multiple transactions. That situation, added to the hardware evolution, have determined the relational databases over the time. One of the different approaches is the actual SAP HANA database. This database focus on the financial and the analytical reporting without the use of the Data Warehouses. But it also provides good capabilities for the operational reporting. That was achieve through the use of a column store structure in main memory. But in order to prepare the data in the database, it holds up the insertion performance. This document studies the possibility to use a buffer in a prototype based in the SAP HANA database architecture, with the goal of improve that performance. In order to compare the impact in the system of the addition of a buffer, multiple approaches has been implemented, tested and carefully compared to each other and also to the original prototype.Grado en Ingeniería Informátic