Mozilla plans to release Electrolysis (multi-process architecture) with Firefox 36
Electrolysis (e10s) is one of the core improvements for Firefox that Mozilla is currently working on. The feature adds so-called multi-process support to Firefox in a way similar to how Chromium-based browsers make use of it already.
A multi-process architecture that separates the browser's core from open websites and plugin contents improves not only the stability of the browser but also the security of it.
This should not be confused with sandboxing though even though Electrolysis is the gateway to make that happen later on.
Mozilla implemented Electrolysis in Nightly channel versions of the Firefox web browser back in February. The implementation was experimental back then and disabled by default.
Tests showed that work needed to be done, especially in regards to stability but also compatibility with add-ons.
Work has continued on Electrolysis and a roadmap was released recently by Mozilla developer Chris Peterson which puts Mozilla's current development and launch plans in regards to the feature on paper.
It needs to be noted that the roadmap is not set in stone and that bumps in the road may delay the project.
Firefox Multi-process architecture roadmap
- July 18, 2014 - Milestone 1: make E10s usable for average Nightly users but is not enabled by default.
- July 21, 2014 - Firefox 34 development begins. Mozilla wants to use the six weeks that follow to get Nightly users and add-on developers to test e10s and especially add-on compatibility.
- September 1, 2014 - Firefox 35 development begins. Mozilla plans to reach Milestone 2 in this development period. When Milestone 2 is reached, Electrolysis is at a point where it can be enabled for Nightly users.
- October 13, 2014 - Firefox 36 development begins. This is the version of the browser where Firefox's multi-process architecture will be moved from channel to channel (Nightly > Aurora > Beta > Stable) so that it is released to the stable channel of the browser on February 16, 2015.
A change in architecture is a major change and one of the consequences of implementing e10s is that there are add-ons that are not compatible with it.
Add-ons that are not compatible right now are among others Adblock Plus, LastPass, RequestPolicy, Greasemonkey, HTTPS Everywhere, BluHell Firewall or Video Download Helper.
Mozilla keeps track of add-on compatibility with e10s on the Are We e10s yet page. Here you find bugs listed that you can follow to find out if progress is being made to make the linked add-on compatible.
Many popular add-ons have not been tested yet, with the page only listing some of them.
Still, it is very likely that most add-ons that are still actively developed will continue to work as developers will fix them to make them compatible with e10s.
Other add-ons, those abandoned by their authors on the other hand may become defunct when e10s gets released to Firefox Stable. (via Sören)
Btw, MSE is enabled in Beta 36.
Firefox multi-process delayed. Firefox 36 not support it, and will be in future.
NOT TRUE!!! e10s was never planned to be released with Firefox36. see Firefox respond:
Sorry to get your hopes up, but we never planned to release e10s with Firefox 36. We’re holding e10s on the nightly channel until it’s stable enough to roll out to aurora. So we disabled e10s about a week before 36 merged to aurora and then we re-enabled it on nightly shortly after the merge. We’ll probably do the same thing when 37 merges to aurora.
We have a set of eight milestones for Electrolysis. Milestones 1 through 3 are finished and we’re working on milestone 4. When milestone 6 is complete, we’ll roll out on aurora. When milestone 8 is complete, we’ll roll out to beta. You can track the progress of each milestone by searching for bugs in bugzilla with a given tracking-e10s value. For example, tracking-e10s=m4+ is for milestone 4 bugs.
-Bill
You are right, and I mentioned that in the article: It needs to be noted that the roadmap is not set in stone and that bumps in the road may delay the project.
Firefox’s official 64bit final version will launch than version 36. yes?
https://support.mozilla.org/en-US/questions/1005200
I know of no plan or even desire to do so.
oh my god :-o still nottttttttttt
@El-D
It isn’t multi threaded. Not for anything substantial. FF only uses one core and as such half/three-quarter or more of processing power is unused. JS or rendering a page hangs the gui, backend CPU spikes hangs page JS and rendering and gui.
“Thus the MSE/DASH spec has been a moving target until about half a year ago, when Firefox finally started working on it.”
Yet it was working for all browsers besided firefox at the start of last autumn. Only firefox lagged behind…
Looks like it was only a moving target in mozilla’s imagination.
“Google even had the gall to force the issue by disabling non-DASH streams”
Naturally. Can’t wait for those who neglect to keep up.
“or make it so nightly users of Firefox can’t fall back to Flash just a few days ago, even though Mozilla still barely has their MSE support working at this point with a number of critical bugs”
Nightly users are devs/testers. It’s to be expected. Plus probably MSE should be finished on the version numbers nightly is on.
“Google seems to think their priorities are everyone else’s, which is why projects like e10s take so much longer than we’d all like.”
Google has no contoll over how sluggishly mozilla works. Like 2-3 people are actively working on MSE. And ~0-3 people were working on electrolysis for years, because mozilla was busy degenerating the UI.
“Not to mention that Google knee-capped Mozilla by first promising to solve web video problems by creating their own video format, WebM, which would replace H.264 as the standard. And what did they do once Mozilla supported WebM? They said “nah, changed our minds, get H.264 working too.”
Hah. Don’t make me laugh. Knee capping… Google tried to popularize webm VP8/Webm but failed. Naturally since AVC is better. Mozilla got on the train because it fit it’s own shallow ideologies. And since everyone kept using AVC…
“And here, also get VP9 working, because want to fragment web video even more.”
The shame of them. Want to improve video quality/compression. We all should use mpeg-1
Someone’s being a zealot…
Does breaking these addons just reflect the price of progress? Or, does it reflect an anti-user-friendly agenda, catering to Google et al who desperately wish (stipulate) that these thorn-in-their-side addons will wither and die?
“Adblock Plus, LastPass, RequestPolicy, Greasemonkey, HTTPS Everywhere, BluHell Firewall or Video Download Helper”
I’m reasonably certain that I will cease using Firefox if/when either RequestPolicy (currently works, but is no longer maintained) or Greasemonkey/Scriptish are rendered inoperable.
@cynical_n_disillusioned
Wow. They’re barely catching up technologically and you’re bitching about it?
FF as it is is inferior to all alternative in many aspects of the backend. Especially with the lack of multi-threadedness and the years of handicap with HTML video tech too…
Chrome was multi-process from the start. IE is multi process for a long while.
MSE/DASH videos work with Chrome since 2012 and will only work in FF 2-3 versions from now.
Addon devs need to keep up. (BTW adblock plus is already claimed to work)
I don’t mind if you want to get on your high horse about this and misinform people, but first off it’s multi-process tabs they lack, not multi-threading. They’ve had multi-threading for years, and it’s been improving with almost every release. Secondly, Google is in control of MSE/DASH and basically all of web video. Thus the MSE/DASH spec has been a moving target until about half a year ago, when Firefox finally started working on it. Google even had the gall to force the issue by disabling non-DASH streams when they feel like it on YouTube, or make it so nightly users of Firefox can’t fall back to Flash just a few days ago, even though Mozilla still barely has their MSE support working at this point with a number of critical bugs. Google seems to think their priorities are everyone else’s, which is why projects like e10s take so much longer than we’d all like. Not to mention that Google knee-capped Mozilla by first promising to solve web video problems by creating their own video format, WebM, which would replace H.264 as the standard. And what did they do once Mozilla supported WebM? They said “nah, changed our minds, get H.264 working too. And here, also get VP9 working, because want to fragment web video even more.” Of course Mozilla had been fighting for a freer codec to be the web standard all along, so it was a horrible about-face for Google to pull.
“Mozilla plans to release Electrolysis with Firefox 36”
Yay! Well, have Electrolysis in FF40.
@martin
Unfortunately Firefox OS seems like the project that might kill Mozilla so I’m less interested in that, I must admit. They’re spending so much effort on that but don’t seem to be going anywhere on Android. At best it will be a 3rd world, non-English product.
it seems like at the rate firefox is losing users that those six people left might give this a try when its ready. firefox developers have lost their way along with more and more addons.
Yet the Firefox users don’t care to help, just bitch and moan, so you Firefox users got exactly what you deserve.
Electrolysis, for me, is just an additional step for Firefox to become disturbingly similar to Chromium/Chrome. I hope that the developers of Pale Moon and some other Firefox forks will never implement Electrolysis.
That’s just incredibly ignorant. You ever wonder why all of Firefox just slows down to a crawl when you open up multiple tabs at the same time? That’s because everything is tied to a single process.
Splitting that up into multiples processes only helps. It helps immensely. Splitting the rendering and UI processes would prevent that kind of stuttering and slow-down from ever happening again.
Do you know if this has any influence on their Android mobile version of Firefox? It’s a very scary situation to see that Mozilla have absolutely zero market share in the mobile space. I don’t think that many people even acknowledge this. For example: http://cdn.arstechnica.net/wp-content/uploads/2014/04/mobile-trends-2014-03.png
Don’t get me wrong, I think mobile is over-rated but to have zero share, and for that crappy native ‘Android Browser’ to have over 20% share, that’s just not good. Mozilla can’t continue to influence standards if they don’t have a presence on mobile. That there’s no real competition on mobiles is also bad news.
Firefox OS seems to support it, don’t think Android does right now though.
Does Electrolysis improve Firefox performance?
The major way this system should improve performance is that when one tab is eating up processing power, it will not affect other tabs or the user interface as much. It will also probably help mitigate the issues 32-bit users have where certain horribly-designed web apps will eat up all of their memory and crash the whole browser, instead of just a tab (we’ll see if it actually helps anything, though). It will also make it far easier to implement a “tab task manager” type of system similar to Chrome, where we can see how much CPU time each tab is eating up, and kill the ones you want to kill right from there.
I have no information about that, may be to early to determine that.
Hi Martin, can you do an article about Firefox performance tweaks that still work?
Let me think about it.
Good, IE **8** has this feature!
That was about the only modern feature it had, and it’s really not all that useful or important anyway compared to the things Firefox had back then. Some people just think this is the second coming when it really isn’t. It will help, but not that much.
Even FIrefox developer said that old addon can run fine in e10s then I cannot trust him
I believe most (or all?) “old” addons which were built with jetpack should continue to work without modification.
Seriously? GM, or GM Port, is somehow causing https://bugzilla.mozilla.org/show_bug.cgi?id=1013744 ? … I don’t believe that I’ve seen any code that does this intentionally nor unintentionally… I did a little further reading in that bug at https://bugzilla.mozilla.org/show_bug.cgi?id=1013744#c1 and it appears that it was a false/mistaken report and someone might want to consider having this corrected in http://arewee10syet.com/
Well, those are some big add-ons for me, but I’m not going to worry about it at this point.
Yeah, I was going to comment on how a few of them seem to have no progress/updates since 2013 on this but considering that there’s still 4-5 builds to be released for the non-beta folk (which would put it toward December of this year likely) it’s not something I’m in a panic over just yet.