Variances and correlations in analyzing the performance of DBMSs with different resources. Breaking points of the graphs

Experiment results

DBMS performance (long moving)

Fig.1. Test VM-1. DBMS performance

Fig.1. Test VM-1. DBMS performance

Fig.2. Test VM-2. DBMS performance

Fig.2. Test VM-2. DBMS performance

Moving variance of DBMS performance

Fig.3. Test VM-1. Sliding dispersion.

Fig.3. Test VM-1. Sliding dispersion.

Fig.4. Test VM-2. Sliding dispersion.

Fig.4. Test VM-2. Sliding dispersion.

Correlation between performance and DBMS expectations

Fig.5. Test VM-1. Correlation between performance and DBMS expectations

Fig.5. Test VM-1. Correlation between performance and DBMS expectations

Fig.6. Test VM-2. Correlation between performance and DBMS expectations.

Fig.6. Test VM-2. Correlation between performance and DBMS expectations.

Results and conclusions

Although the performance indicators for the test VMs differ by orders of magnitude, nevertheless, common features are noticeable:

  1. The presence of a breaking point in the moving dispersion graph. This is especially noticeable for a less productive DBMS.

  2. The presence of a breaking point in the correlation graph between performance and DBMS expectations. Although the nature of the graph is quite different.

  3. A more detailed analysis of the DBMS expectations is needed.

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *