*Parallel testing moving forward:

*Improve reliability:
	* Additional runs for edge case testing
		* Duplicate runs for each service
	* Run in idle case on the cloud with extra cloud capacity we need
		* Use new zuul scheduluer to schedule jobs when possible
	* Developer triggered ability to run 10x jobs

*Further speed improvements
	* Increase number of test runners:
		* Not necessarily good for speed, but may help shake out more race bugs
		* up to 8 as a first step
		* run in "idle cloud" at next layer up of stress
	* eliminate setupClass?
	* Distributed testing?

*Scheduler changes
	* Preseed time data? Probably not worth putting time into, however if you do do it put it in a repo behind http so that it can be centralized
	* Set number of workers

*Logging
	* Howto make parallel test logs easier to read.
		* Add test class name?

*Testr UI vs Tempest UI
	* Exclude lists in nose (either a file or cli argument
	* Update testrepository bug with idea of using rsync as a base grammar for test exclusion

*Documentation:
	* Are there gaps in tempest docs around issues with parallel testing
	* What open questions need to be answered in docs?


*Drop tempest python 2.6 support:
	* Remove all nose hacks in the code
	* Delete nose, use unittest2 with testr/testtools and everything *should* just work (tm)