Automated storage benchmarks for embedded storage
Define benchmark(s) for SD and eMMC storage for root-filesystem usage, possibly based on IOZone. These must account for wear effects, which cause system performance to decrease as the cumulative number of write operations increases. Such benchmarks also must be automated.
Blueprint information
- Status:
- Complete
- Approver:
- Paul McKenney
- Priority:
- Medium
- Drafter:
- None
- Direction:
- Approved
- Assignee:
- None
- Definition:
- Obsolete
- Series goal:
- Accepted for 11.05
- Implementation:
-
Started
- Milestone target:
-
backlog
- Started by
- Paul McKenney
- Completed by
- Deepak Saxena
Related branches
Related bugs
Sprints
Whiteboard
Using and existing benchmarking framework such as IOZone, create a benchmark for a root-filesystem workload.
Implement or use existing benchmarking that runs on a block device level to avoid the affect of the page cache. This will give more precise numbers when optimizing the low level memory frameworks such as the mmc framework and driver.
Using existing tooling, automate extracting from system logs the storage-wait-time contribution to system boot.
Analyze results, accounting for wear effects (e.g., fudge factor on results based on cumulative number of writes).
Script running benchmark on various rootable filesystems (for example, ext2/3/4 and/or btrfs).
Test robustness. Reset the power while performing write operations, reboot and verify the filesystem.
Work Items
Dependency tree
* Blueprints in grey have been implemented.
