Loading...
Please wait, while we are loading the content...
Similar Documents
When Logs Become Big Data
| Content Provider | Semantic Scholar |
|---|---|
| Author | Iversen, Morten Aursand |
| Copyright Year | 2015 |
| Abstract | As we move into the era of Cloud Computing and the Internet of Things, an increasing amount of devices are connected to our networks and this is expected to be doubled in the next five years. This results in large amounts of logs, sensor data and other metrics that has to be stored and analyzed. In this project three databases are compared from a log analytics viewpoint. These databases are Cassandra, Elasticsearch and PostgreSQL. Experiments are designed and run to test the general performance of the databases with write and read operations, in addition to some experiments that are designed to look like normal use cases from log analytics. Some of the experiments are repeated in an Elasticsearch cluster of varying sizes to see how this influences the performance. The results indicate that all the databases get quite similar results in the general performance tests, but that Cassandra does very poorly in the use cases that try to simulate log analytics. It is concluded that PostgreSQL and Elasticsearch are both good options. And the results from the clustering experiment indicate that Elasticsearch would scale up very well, meaning that it is well prepared for future needs. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | https://www.duo.uio.no/bitstream/handle/10852/45115/Iversen-Master.pdf?isAllowed=y&sequence=1 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
