Timeline
Week ending Nov 10:
- Black box comparison of image files
Get detailed information on number of page matches
If very little common pages are
identified the reason needs to be identified. (Possible reasons
could be that the pages of a
process are differently distributed in a different way)
- Experiment with stop/copy migration (rather than live migration) and gather data
Week ending Nov 17
- If the black box comparison did not give good
results, reading the kernel data structures should be considered to
identify reason
- Figure out how to use ballooning driver to release the free memory pages to Xen when not migrating
Week ending Nov 24
- Integrate ballooning into current migration framework
- Continue work on pre-deployment
Week ending Dec 01
- Sensitivity study on time migration
- Continue work on pre-deployment/ integration into migration framework if pre-deployment is found feasible