Thread by @GrapheneOS

microG does NOT provide an open source implementation of Play services. Apps using Play services integrate the closed source Google libraries which are fully capable of contacting Google services regardless of whether Play services is present, and also do so in practice too.

The implementation of the services is also still closed source. It's only an open source middleman and the Google libraries are perfectly capable of communicating with the Google services without it. It doesn't provide better privacy or security and doesn't make Play open source.

It's not as secure as the official Play services implementation and doesn't provide comparable privacy for user data. It leaks user data across apps, doesn't properly enforce the API security model and doesn't provide comparable transport security via pinning and other measures.

It has always been possible to install microG on GrapheneOS. However, we have never been willing to make the security sacrifices necessary for it to work properly. It requires bypassing security checks in apps for the official Play services signatures without having parity.

We of course also weren't willing to integrate Play services into the OS even as an optional feature. The standard integration requires massively invasive access and privileges in the OS. It expects extensive access to internal privileged APIs and special SELinux MAC/MLS policy.

Play services is extremely problematic because it can do far more than other apps and services. It doesn't play on a level playing field. Modern Android has a great application sandbox and strong privacy and security. Play services expects to be built into the OS, bypassing that.

Sacrifices required to integrate either microG and Play services are both completely unacceptable. microG is also hardly a real alternative. It implements a tiny subset of the APIs. It's missing most of the functionality. It can't provide the broad app compatibility we needed.

GrapheneOS was founded in 2014. We've also wanted broad app compatibility. There was never previously a way for us to do that meeting our requirements.

Funding development of grapheneos.org/usage#sandboxe… has provided what we needed: a no compromises approach to Play compatibility.

Installing Play services on GrapheneOS was always possible but as a fully sandboxed app with no special access or privileges, it couldn't work at all. It would simply crash repeatedly as it tried to use the massive amount of privileged functionality that it expects from the OS.

It took us years to realize that there was a highly maintainable approach without privacy and security sacrifices: teaching Play services to work as it should have functioned in the first place. Fully sandboxed without special access/privileges like every other library/service.

Installing Play services on GrapheneOS within the user or work profile(s) of your choice provides broad compatibility with apps in those profiles using Google's libraries. It provides ZERO additional access or privileges to the Play services code in each of the apps using it.

You don't have to grant any permissions or other access to Play services in order to have broad app compatibility. You don't have to sign into a Google account, but you can do that if you want and you can use a throwaway account just like Aurora Store does for you automatically.

Our approach isn't a compromise. GrapheneOS doesn't include Play services. It doesn't offer any special access or privileges to Play services if users install apps depending on it or the Play services apps themselves. It works like any other app or service. Same rules and limits.

Implementation already provides most Play services functionality. Went from a concept to highly functional in 2 months. This is the total code size across the 3 repositories with changes:

47 files changed, 1556 insertions(+), 11 deletions(-)

Highly auditable and maintainable.

It took us 7 years of putting substantial thought into this to settle on the right approach. It's understandable that stakeholder in legacy alternatives don't see it our way yet. It's unfortunate that they've chosen to spread misinformation about our approach to this though.

We're more than happy to explain the extreme flaws with every other approach that we've considered, and why this makes the most technical sense.

If microG had any serious potential of broad app compatibility, maintainability and security we would have integrated it YEARS ago...

Bundling an existing solution is the easy path. It'd literally take 5 minutes to integrate microG.

GrapheneOS is not about taking the easy path but a proper approach able to provide our users with a high level of usability, privacy and security. Shortcuts aren't our thing.

Same reason we don't integrate F-Droid but rather are making a new app repository system able to meet the usability, privacy and security requirements needed for GrapheneOS over the long term.

https://twitter.com/GrapheneOS/status/1415346236673175557

We're investing in making great tech, not marketing/branding.

It's the same reason we made hardened_malloc rather than sticking to bolting on far weaker mitigations to existing malloc designs.

Often means taking longer to get there, but we eventually do. Web installer is a nice example. Is there an easier OS to install than GrapheneOS now?

Our community sees the value in our approach and GrapheneOS has a quickly growing full-time development team funded entirely by donations. No business model and no reason for one. It's not funded by grants either. Bureaucrats have never seen the value of work. We don't need them.