Android 14 introduces cellular connectivity security features
security.googleblog.com
security.googleblog.com
There's some regional carriers in rural areas that offer the only coverage available. Like Commnet Wireless (2). These are few and far between and usually they have deployed 3G to their whole footprint. The Big Three are building out native coverage to overlap with them. But by Murphy's Law someone with an Android 14 phone is going to discover that they can't call anything but 911. Ideally there would be a button prompt enabled in No Service situations to re-enable 2G. FCC rules mandate that cellphones must support fallback to null cipher if that's what's needed to connect an emergency call.
(1) https://www.t-mobile.com/support/coverage/t-mobile-network-e...
(2) https://www.cellularmaps.com/regional-carriers/commnet-wirel...
I was somewhere between Hazen and Bismarck recently and had no signal.
the remote areas with less connectivity are probably places without deep-pocketed corporate customers that would justify the expense of running fiber. I suspect large swaths of the Alaskan interior fall under that description.
0: https://alaskapublic.org/2017/12/05/new-fiber-optic-cable-sy...
Starlink has to travel minimum of 550km twice + distance from ground station to destination server. For remote locations, you generally also have an additional satellite to satellite relay distance as well. So realistically say 1500km minimum, likely way higher. Speed of light time on that distance is ~5ms, and that doesn’t account for latency losses at each hop and also remember that’s one way time.
There’s a reason microwave is still used, even sometimes over fiber, where latency matters.
So it's about +10ms to your ping for nearby destinations, less (and potentially even faster) for more distant destinations. That's really not enough to worry about in most contexts; even for FPS gaming, unless you're a pro 10ms isn't a huge difference maker.
> There’s a reason microwave is still used, even sometimes over fiber, where latency matters.
Only for HFT as far as I know; latency mattering that much is a tiny niche.
You obviously live in a tech bubble. Remote industrial machine control definitely has concerns around guaranteed latency.
Oh, so you are finally conceding that Starlink doesn’t fit every use case? Because Starlink definitely goes down during a storm occasionally but a dedicated microwave link with high gain parabolic dishes won’t in the same conditions.
> I very much doubt there are any cases for remote industrial machine control where microwaves are the better option.
Oil and gas industry would beg to differ.
[1] https://www.reddit.com/r/Starlink/comments/yllqk6/space_lase... [2] https://www.highnorthnews.com/en/elon-musks-starlink-expands... [3] https://www.starlink.com/map
I was thinking that they were way overdue to replace those microwave dishes with a Starlink terminal.
Deadhorse also has a fully functioning commercial airport with Alaska Air flights that are "free" for the workers there.
neat, I didn't realize the US had anything like 'company towns' anymore, hopefully they avoid the paternalism stuff that ruined company towns in the previous centuries.
I've lived in the heart of a 2M+ city, and the skyscrapers made reception so difficult that AT&T gave me a free pico cell for my apartment.
There are places in California that only have 3G, as of last month when I visited one.
It's a different world outside the Bay Area bubble.
I will never understand how the powers that be decide where to deploy coverage.
That didn't happen for a solid 30 or 40 miles, because none of us had signal. And that was on an interstate.
TLDR; Cellular providers make money by selling you new equipment, and they will claim a device is unsupported even when it's not true.
A US-only analysis of this seems not especially useful, since Android is used worldwide. If anything it is more popular outside of the US than inside of the US, making US-based analysis even less illuminating.
Getting rid of null ciphers is good though. It would be nice to also refuse weak, export, etc ciphers.
I don't think this setting does what you think it does. The description under this option has a big caveat: "For emergency calls, 2G is always allowed". So even when disabled, the phone can still use 2G networks.
It sounds like this new option is to actually disable all 2G functionality.
Previously (e.g. Android 13) if your cellular provider said you can choose "No" then there's a slider in your Settings so you can choose, otherwise it's greyed out and tells you "Name of Your Provider requires 2G to be available".
On corporate devices it's a pain to have settings that your users can screw up, and so it's nice when Android (or Apple for iOS) makes the setting something you can choose to lock down instead.
# Edited to correct text of fixed message when your cellular provider has decided you mustn't switch off 2G.
There is also nothing the OS can do to shut down 2G in most cases; the best it can do is ask nicely that the baseband does not use it. So the "shut it all down" theory is also not correct.
It’s probably a legal requirement that if 2G is supported emergency calls on 2G must be also supported. Maybe there are some LTE-3G dualmode phones without GSM in the world, if that’s what you must have?
I'm pretty sure that it was intended that the OS UI would show you when your connection is unencrypted, but none of them do because that was undesired by state actors.
Also, even if encryption is enabled it's only for the radio part of the data transmission, not handset -> handset. Otherwise you would not be able to make calls to landlines, so isn't it already trivial for a Network Operators to decrypt your raw data? It would help for scenarios like an embassy mounting a fake base station to grab data about protestors outside it, I suppose.
Also, how can they tell if the encryption key is weakened by setting lots of bits to zero, like was done in the original version of GSM?
I would be happy to have that on a GrapheneOS phone for example, if I hadn't went with Apple.
You cannot apply Hanlon's Razor to megacorporations.
I wouldn't say malice but some of their practices are hostile.
Fix the issue now that GSM calls are rare...
This is great if the phone decides 3G or 2G connectivity is better, but I know the 4G network is faster (still slow). A downside is that if the 4G network is completely overwhelmed (e.g. on a festival), the phone might not receive phone calls or sms it'd receive if it could switch to 2G (happened to me. The phone had 4G connectivity, but SMS didn't work without allowing 2G).
i was running various jelly bean custom roms and band selection was a common feature iirc
2G is long dead.
I'll be honest. The stuff in this article is good, if a little underwhelming, but I feel a large amount of distrust for Google nowadays, to the point where what would've felt like unnecessary pessimism now feels only rational to me.
Ever since Google dropped WEI into our lives, I feel like they should not be allowed to be a part of any security efforts in any standards body or ecosystem. How long until carriers try to limit devices that don't support Google Play or Apple remote attestation of some kind?
I don't know what to think or do anymore.
Even the parts that do good feed the parts that do bad.
In the end there's a net result. Doesn't matter if it "put out a good backup app", if they also got to own the web or build Skynet.
There is, but the financial net result is only known with hindsight. Exactly which actions boost the net result in a particular time frame always remains uncertain.
Also, the people who work for corporations have their own goals and net results, such as promoting their own careers or self esteem.
Shareholders may wish for a single minded, deterministic, profit seeking machine but that is not a possibility. Capitalism just provides vague, incomplete and even contradictory incentives.
I mean a net result regarding the impact it has on society.
>Shareholders may wish for a single minded, deterministic, profit seeking machine but that is not a possibility.
Yes, but my point is not based on some total purity of absolute profit seeking.
It's enough that we can observe for 20 plus years now actions with harmful intentions and harmful effects, primarily motivated and explained by profit seeking.
It doesn't have to be Google's (or any companys or organizations) absolute and only priority, nor everybody has to be single-mindedly going for it all of the time, Google doesn't even have to be totally effective in this goal for them to perform profit-increase-oriented harmful actions.
I mean, it's such a pedantic point! What you wrote is common sense, and it wasn't contended. I didn't paint them as pure-profit-seeking-pure-evil-nothing-else.
Just as an actor that causes harm in pursuit of its profit-seeking.
Yes, because a company can be a negative actor driven by profit-seeking only if it is "controlled on the micro-level by some all-knowing AI with the single goal of profit seeking, and it can ultimately align every one of their goals to profit".
To think otherwise would be naive /s
Thanks for the nuance!
So it's difficult to trust them with anything. There are plenty of great people at Google, but they're not able to stop the awful things Google does, so why does it matter?
Exactly zero folks are liable to notice that you have a subtle nuanced analysis regarding their strategic direction with a feature. It would be like you telling a star about how you feel about how they wore their hair at a show 3 weeks ago. Your opinion is either invisible or irrelevant.
I know I'm biased in my opinion on WEI, my initial reaction was extremely negative.
That said, I'm willing to be open minded here. As a user, how does WEI benefit me?
I can easily see how WEI benefit's advertising companies like Google and how it benefits advertisers that want to track their campaign's effectiveness. What I can't see is how it benefits a user of the web.
I don't care if I'm responding to a bot if the conversation is meaningful, nor do I care if a bot posted some content to whatever content feed I'm currently browsing because I like to see content and it doesn't matter if a bot or human posted it, to me.
What is the benefit in "improving websites' trust in me" to me as a user? As a web developer, I do not trust my users nor would I ever trust them. This is a core principal in application development.
The "give up your freedom to prevent crime" argument isn't new or compelling. I'd rather keep what little privacy I have and put up with spammers.
As for captchas and logins, I'm not sure that'll change much. Websites still want to spam us so they'll keep asking for our email addresses, and companies still have AI to train so they still have to force us to identify buses and crosswalks. I also don't care for the idea of participating in a protection racket. "Just give us all your data and we'll stop annoying you by demanding it" isn't much better than "just pay us and we'll stop trashing your property"
Alright, you got me. There is a tiny, minor, insignificant benefit. This is not worth giving up any control of my own PC.
> You'll see less spam in website comments.
Highly unlikely. This will be defeated by spammers just like every other attempt to stop them is.
> Fewer websites will be behind a login wall.
Not just highly unlikely, but absolutely not going to happen. No site that currently requires login will change to not require login. No site will give up that user data.
Wait, so no Google or Apple employees involved in any standards body security efforts. What about TPM? Better ban employees from Intel, AMD, Qualcomm, Microsoft...who's left?
I mean, that's a take, but it seems like really the take away is that we should be skeptical of company motivations and security issues in standards bodies should be dealt with transparently, which all seems like a good take?
I truly believe Google is off on their own island of being impossible to trust in their current state.
Especially not in a world where W3C gleefully standardizes literal DRM.
Number is irrelevant - power is.
>seems like the regulators should look at Apple first, not Google.
If some one person company proposed this instead, would you want regulators to look at them first?
To put it simply - Apple does not provide OS to smartphones and PCs, it manufactures them. So as long as they do not produce 50% of PCs or 50% of phones or have Microsoft agreeing with their proposal then they are lesser threat (for now).
It's like with Monopolies and Cartels - power over market should bring relative response from regulators over company - especially if that company tries to create proposal which invades users privacy.
>So as long as they do not produce 50% of PCs or 50% of phones
Apple's market share in the US smartphone market is at almost 58%.
>It's like with Monopolies and Cartels
Yes, and I'm not saying Google shouldn't be looked into if necessary, but Apple has a 100% monopoly on iPhones unlike Google on Android and PC where you can switch browser.
If Apple is stopped, there are no reasons to go after Google, as it would also be illegal for them to do the same. Two birds with one stone – and Apple definitely also needs to be stopped.
And less than 30% worldwide.
> and PC where you can switch browser.
Under condition that you are not using anything other than Windows/ChromeOS - then no matter how much you switch the browser the OS stack will still not be certified - like DRM always fails to work on my end.
>If Apple is stopped, there are no reasons to go after Google, as it would also be illegal for them to do the same.
You are missing the point for me - Apple is just somewhat big¹ hardware company with niche¹ system. If you want to fight them because they have enormous power on your market - power to you, but from global perspective "no reasons to go after Google" is clearly misguided.
¹from my country perspective
P.S.
It isn't worse - both are terrible and both should ultimately be scrapped. Just because one company invents invasive bs doesn't give rest right to continue with that.
Before someone says Apple has 50% of phones - no it does not. Based on worldwide data not US one.
Intent.
to say you don't trust google to be part of any security efforts because they tried to put security in a place you don't want it is silly. you're arguing the slippery slope fallacy here, there's no reason to think that carriers would even want any sort of device attestation, or be legally allowed to do that under the terms of their spectrum licenses.
There are quite a few of those "dual use" features where you have a primary purpose of protecting some oligopolist's business model but the incidental side-effects range from beneficial to catastrophic in terms of security for users.
Take Apple's side-loading ban for instance. It makes obvious sense for Apple's own interests. Beyond this primary purpose, it protects people against getting tricked into installing malware but it also poses a threat to human rights and freedom of speech as it gives whoever is in power central control over the software people can use to protect themselves.
Google's WEI is another such dual use technology that can be portrayed in many different ways, because it really does have different consequences for different people at different times.
>there's no reason to think that carriers would even want any sort of device attestation
With the bargaining power that Google has in the mobile space, I strongly suspect they could find a way to entice carriers into moving towards enforcing device attestation. And I don't just mean forcing devices to attest; I mean things like, Hmmm, are you sure you want these untrusted Android devices on your RCS services?
> or be legally allowed to do that under the terms of their spectrum licenses.
I don't really like relying on hoping the laws will help, but it would be nice to know for sure if they do. Unfortunately, I am assuming you aren't actually positive.
Google Fi definitely favors Google devices heavily and there is no supported way to use a non-data-only device without an Android phone with GMS.
Nobody said you did. Normally, crummy security(security flaws) is what disqualifies someone from security and not functioning security. You're comment implied WEI would work as imagined and you remarked that should disqualify Google from security. That's a bit ass-backwards, disqualifying someone because they didn't introduce security flaws.
> The problem is that it's a terrible, user-hostile, open-web-hostile approach to security that shouldn't have even been proposed.
But it isn't a security flaw. And no, it isn't a terrible, user-hostile, open-web-hostile approach to security. It's just a gleam in someone's eye who's stated goal is to not create something which is a terrible, user-hostile, open-web-hostile security approach but some additional signal which websites can try to use to filter out malicious traffic. Why do WEI critics always come off as sounding like the "think of children" anti-encryption fear mongers? At least wait until it becomes something first.
> With the bargaining power that Google has in the mobile space, I strongly suspect they could find a way to entice carriers into moving towards enforcing device attestation. And I don't just mean forcing devices to attest; I mean things like, Hmmm, are you sure you want these untrusted Android devices on your RCS services?
With the bargaining power that Apple has in the mobile space, I strongly suspect carriers can tell Google to pound sand. Unless Google is implementing something Apple has already, I doubt that will happen.
Yes, developing bad, user-hostile, open-ecosystem-hostile security standards should indeed bar you from security standards work even if the prison you've built for end users is maximally secure.
Actually they did introduce a giant security flaw: https://news.ycombinator.com/item?id=36985317
Because of this I propose that "the users should require that server attest it, environment and content with respectable attester before it provide any other data to the user - user should get signal if server operates on up-to date unmodified stack and did not modify server environment in any way to track, monitor, gather (steal) user's data, nor is compromised." How about that?
>"the the bargaining power that Apple has in the mobile space"
Like 30% (and loosing) of worldwide phones? Think globally not US. And we shouldn't forget that Apple may loose EU market altogether because of it inflexibility when comes to battery replaceability.
>Why do WEI critics always come off as sounding like the "think of children" anti-encryption fear mongers?
No we do not. In fact many just want opposite - for encryption not to be used as power over common people but by the people.
> "At least wait until it becomes something first."
As Google implements it before even finalizing proposal and banned all comments and pulls - I will not.
> It's just a gleam in someone's eye who's stated goal is to not create something which is a terrible, user-hostile, open-web-hostile security approach but some additional signal which websites can try to use to filter out malicious traffic.
To put it by analogy from Car owning company: "We propose that any package delivery must be delivered in secure and attested way - so we propose mechanism that will provide way for smaller companies fly drone before you get your package. We will not allow them to sell your photos - pinky promise."
This is more or less how I feel about Google promises.
If you propose something on global - you should take all the time to think about ways that it can be mismanaged before you even propose it, If you trivialize concerns then you are as guilty when they inevitably happen.
>some additional signal which websites can try to use to filter out malicious traffic.
"The solution to the surveillance economy seems to be more surveillance."
This is quite good quote - because if you think for second about this phrase it should give you good idea what You can get…
WEI will be the ālea of Internet.
Excluding Apple and Google, the remaining bodies are MS, Amazon and Facebook which presence is close to non-existent in the mobile OS market. Good luck with them?
Meanwhile, the mobile ecosystem is still rife with many other avenues - your MVNO, fractured Ma Bell, Play Services still has outsized privileges on standard Android, most apps aren't E2EE (despite the article's bastardized use of the term), etc. It's just this boring corporate security narrative where we're supposed to continue ignoring the 800lb gorillas selling our personal information into countless surveillance databases and focus on how they're closing down possible independent attackers.
What would be newsworthy is if they were even talking about real security - libre baseband, mitigating protocol identifiers (eg IMEI) that allow for pervasive location tracking, etc.
The press release [1] doesn't mention 2G directly but only "safer wireless connectivity defaults" but FastCo [2] is more direct "with iOS 17, Apple is not only beefing up Lockdown Mode (by blocking the iPhone from connecting to 2G cellular networks and from auto-joining insecure wireless networks) but bringing Lockdown Mode to the Apple Watch".
[1]: https://www.apple.com/newsroom/2023/06/apple-announces-power... [2]: https://www.fastcompany.com/90904197/apple-ios-17-craig-fede...
I like some of the settings of lockdown mode, and I wish you could enable them one by one. Enabling the entire lockdown mode is a bit too much
If not, it’s way less interesting.
Missed opportunity to call them Mal-in-the-Middle attacks.