Attendees
| Sl. No. | Topic | Presenter | Notes |
|---|---|---|---|
| 1 | Project Repo Status | Repo is ready, we can start submitting patches. https://github.com/opnfv/thoth Submitted research studies. | |
| 2 | Failure emulation work status | Experimenting on Pod18 - 30, 40, 50, 80... Pods.. Will have to increase to 100. Haven't started with "Openstack pods" yet. similar tool with stress-ng is almost ready - will test it next week. - https://github.com/ColinIanKing/stress-ng | |
| 3 | Data Extraction Tool Status | Completed implementing and testing for Prometheus (who is receiving metrics from collectd) data extraction. | |
| 4 | Demo of Failure Prediction Models | Migration to Intel Pod12. Review of various data models. Based on the RMSE tests, CNN seems to be the best option. Compared single step and multiple step training, got similar results. Data was based on 1 minute of traffic. Link to Slides to be added to the notes. Presentation: https://docs.google.com/presentation/d/1awXXzH5piKg0fQiYgyqpMh6xXzBAlJzE/edit |