Electrolysis/Firefox: Difference between revisions
| Line 24: | Line 24: | ||
* Security improvements are a non-goal for e10s. We do believe that this project will net some security gains but it's not an explicit goal. In general the app should be no less secure than Firefox 4 but there is no additional security bar. | * Security improvements are a non-goal for e10s. We do believe that this project will net some security gains but it's not an explicit goal. In general the app should be no less secure than Firefox 4 but there is no additional security bar. | ||
* There are 3 areas that are not really addressed by | * There are 3 areas that are not really addressed by e10s but could possibly be effected. They might improve or they might degrade when we turn on e10s. For this project we would want the following measurements to be equal to the Firefox 4 performance. | ||
** In page responsiveness | ** In page responsiveness | ||
** Page load time | ** Page load time | ||
Revision as of 21:54, 1 November 2010
Background
This is the home page for all things relating to getting Firefox up and running with Electrolysis. The Mozilla platform will use separate processes to display the browser UI, web content, and plugins. The working name for this project is Electrolysis, sometimes shortened to e10s.
Status
We are currently in the planning phase.
Goals
There are number of things we believe the e10s project will give us (ordered in relative priority).
- Better application UI responsiveness.
- Improved performance, especially on multi-core machines.
- Better memory core stats.
- Crash resilience. A web content crash doesn't take down the entire browser.
- Preparation for sandboxing. Web content can be sandboxed more cleanly.
- Better visibility into how resources are being used.
The #1 driver for e10s is to improve the application responsiveness, or the perception thereof.
Non Goals
- Security improvements are a non-goal for e10s. We do believe that this project will net some security gains but it's not an explicit goal. In general the app should be no less secure than Firefox 4 but there is no additional security bar.
- There are 3 areas that are not really addressed by e10s but could possibly be effected. They might improve or they might degrade when we turn on e10s. For this project we would want the following measurements to be equal to the Firefox 4 performance.
- In page responsiveness
- Page load time
- Startup time
Targets
We need to define specific targets in order to quantify our goals.
<insert targets here>
We also understand there will be some costs to certain areas in performance and memory consumption.
<acceptable tradeoffs>
<areas where perf can't degrade>
Implementation
The implementation path will take a phased approach. Beyond Phase 1, we have not fully planned this out so more detail will be added as we make progress.
- Phase 1: Get Firefox up and running with e10s turned on.
- Phase 2: Cleanup what was done after Phase 1.
- Phase 3: Make the browser work properly. Start to think about parallel projects, testing, metrics.
- Phase 4: ...
- Phase 5: ...
Questions
There are some big open questions around the implementation details and approach. These are not fully formed but we are keeping track of them since more stuff will come up as we complete Phase 1.
I have put together a page with all the musings