MÖJLIGHETERNA MED CLUSTERED COLUMNSTORE INDEXES & IN-MEMORY OLTP
Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
Clustered Columnstore Indexes (CCI) and In-Memory Online Transaction Processing (OLTP) are two new techniques that were introduced with Microsoft’s new database management system SQL Server 2014.
The thesis analyzes these two techniques, describes the theory behind them and also the background information that needs to be known to understand more of the new techniques. Performance tests of the new techniques have been performed to see how much faster response time we can get and less storage we need to use where the result is represented by performance diagrams and a compression diagram. This thesis is a cooperative work with a company called Trimma and the result is used as support to decide if these new techniques could be of advantage for the company.
The test consists of five test queries with different purposes. The result is illustrated by diagrams that show the result and another diagram that shows how much the data gets compressed with the different techniques. CCI did get back response results faster in all tests except from one and In-Memory OLTP was the table with the slowest response time in all tests. Regarding the compression was CCI, as expected, the technique that had the highest level of compression compared to the other two techniques and In-Memory OLTP had the lowest level of compression according to how much space the actual data needs on the disk compared to the other two techniques.
The conclusion of this is that CCI saves both time and space on the disk and that In-Memory OLTP needs a lot of space both on the disk and in RAM. Aside from this we need to consider the fact that In-Memory OLTP wasn’t tested under optimal conditions, which should be an environment with high activity and many users. In-Memory OLTP can probably perform better and get faster response time under such conditions.
Place, publisher, year, edition, pages
2014. , 39 p.
, UMNAD, 993
Engineering and Technology
IdentifiersURN: urn:nbn:se:umu:diva-92831OAI: oai:DiVA.org:umu-92831DiVA: diva2:743785
Bachelor of Science Programme in Computing Science