[HN Gopher] Google dragged to UK watchdog over Chrome's upcoming...
       ___________________________________________________________________
        
       Google dragged to UK watchdog over Chrome's upcoming IP address
       cloaking
        
       Author : Beggers1960
       Score  : 91 points
       Date   : 2023-11-12 15:54 UTC (7 hours ago)
        
 (HTM) web link (www.theregister.com)
 (TXT) w3m dump (www.theregister.com)
        
       | cynusx wrote:
       | Wouldn't that mean that all your web traffic is going through a
       | google server?
        
         | andylynch wrote:
         | Yes. It's like a Googled version of Apple's Private Relay.
        
         | layer8 wrote:
         | Yes, but it's end-to-end-encrypted.
         | 
         | See https://github.com/GoogleChrome/ip-protection/issues/10 for
         | example.
         | 
         | Of course, Google can conceivably backdoor Chrome, and then the
         | exfiltration of data wouldn't be obvious from the client-side
         | traffic.
        
           | quenix wrote:
           | If we assume Google to backdoor Chrome, the whole IP
           | protection talk is irrelevant. They could backdoor it without
           | needing that.
        
             | layer8 wrote:
             | Chrome would need to contact Google servers to send the
             | collected data. Such unexpected traffic would be more
             | easily detectable without IP Protection. With IP
             | Protection, Chrome is sending data to the Google proxy all
             | the time, so it would be harder to detect extra data sent
             | that way.
        
             | onedognight wrote:
             | The CA system and hence https/TLS is already backdoored, by
             | construction. Any of root CA's that your browser trusts can
             | man-in--the-middle you unless you keep track of and pin
             | certificates like Google does. These roots CAs are run by
             | random companies and governments around the world. All of
             | them have equal precedence in browsers.
        
               | layer8 wrote:
               | It's not that easy. They would also need to MITM the
               | actual traffic, or the DNS.
        
               | basique wrote:
               | Any fake certificates the malicious CA signs will have to
               | be written to a Certificate Transparency log, which would
               | make noticing the fact that the CA has done something
               | like this easy.
        
         | JimDabell wrote:
         | From the third paragraph of the article:
         | 
         | > It's designed to run Chrome browser connections through two
         | proxies, one operated by Google and one operated by a third-
         | party (eg, Cloudflare), so that the true public IP address of
         | the user is obscured, hopefully thwarting attempts to track
         | them around the web using that address.
        
           | justinclift wrote:
           | Reading the front page the of GitHub repo with the Chrome "IP
           | Protection" source code, it has this worrying statement:
           | We are considering using 2 hops for improved privacy. A
           | second proxy would be run by         an external CDN ...
           | 
           | https://github.com/GoogleChrome/ip-protection
           | 
           | "Considering" means "we're thinking about it, but it's in no
           | way final".
           | 
           | Sounds like Google could implement a proxy solution but
           | decide the 2nd hop for improved privacy isn't needed after
           | all.
           | 
           | Or make it optional, defaulting to OFF, etc.
        
             | buremba wrote:
             | ^ This is an important "detail".
        
           | eli wrote:
           | In other words what apple has been doing for years now
        
           | AnthonyMouse wrote:
           | Wouldn't this be useful as a way to avoid getting blocked or
           | having to do a thousand CAPTCHAs when using Tor? Access the
           | relay via Tor, the relay doesn't know who you are because of
           | Tor, the website sees you as someone using the relay which is
           | too many ordinary people to block.
        
             | mindslight wrote:
             | Yes, but such proxies will likely require authentication
             | with a Google account, and Google will keep an audit log to
             | prevent "abuse". So at best it works out for maintaining
             | somewhat persistent nyms for well-scoped purposes (that you
             | make with a burner SIM or buy from someone who did), but
             | not as some panacea for Internet privacy.
             | 
             | This is of course assuming that the roadmap for these
             | proxies isn't tied into their push for remote attestation
             | ala "Safety" Net and WEI, in which case your "nym" becomes
             | the entire device.
        
         | LinuxBender wrote:
         | _Wouldn 't that mean that all your web traffic is going through
         | a google server?_
         | 
         | Yup just like most people using Android have their text
         | messages routing through Google. Some don't even realize this.
        
           | Andrex wrote:
           | RCS E2EE and Group E2EE are live which makes this a less
           | dangerous attack surface / a less useful ad data reservoir.
           | 
           | Google Voice users like myself are SOL but I've never had
           | pretenses that my messages there are private. They're
           | probably stored in plaintext even, there's a lot of 15+ year
           | old GrandCentral cruft underpinning Voice even to this day
           | (speaking only as a user).
        
       | diroussel wrote:
       | Combined with Privacy Sandbox [0], this means the browser can
       | track you directly without cookies, but websites can't track you
       | via IP address. So it has the effect of centralising the tracking
       | to google only.
       | 
       | I haven't used Chrome since this rolled out, since I didn't see a
       | way to object to the new tracking. This is what the linked
       | article [0] says:
       | 
       | > It's unclear if toggling these features off will stop Chrome
       | from collecting these data altogether, or if it just won't share
       | the data with advertisers.
       | 
       | [0] https://theconversation.com/google-chrome-just-rolled-
       | out-a-...
       | 
       | [1] https://developer.chrome.com/en/blog/shipping-privacy-
       | sandbo...
        
         | jwells89 wrote:
         | Depending on the user's aims, that could still be a net win...
         | limiting tracking to a single party versus N parties. Obviously
         | not ideal as no tracking at all though.
        
           | mindslight wrote:
           | Normalizing VPN/proxy use is a fantastic development for the
           | entire market. Sure, I can see a future where websites only
           | allow traffic from naive connections, Apple proxy, or Google
           | proxy. But I can just as likely see a future where many more
           | "normie"-appearing providers spring up and website managers
           | move on to some other feel-good check box besides IP
           | discrimination.
        
           | pixl97 wrote:
           | The particular issue here is where many people have the same
           | complaint about Cloudflare. You're concentrating that single
           | party to a single provider for a massive number of people.
           | Hence now governments have a one stop shop for all the
           | information they need.
        
           | kevingadd wrote:
           | That's assuming you can trust the single party not to pick
           | winners to share data with and losers to keep out.
           | 
           | With Apple we know they have a bunch of revenue streams based
           | on selling actual products to users, so there's much less of
           | a hazard there. With Google... who knows what they'll do to
           | juice their next earnings report?
        
           | halJordan wrote:
           | It's not meaningfully different. Nobody actually wants the
           | data. They want to do stuff with the data and take actions
           | based on get data. Google will collect all the info and then
           | process it any way the 1st party asks. And the 1st party will
           | still take the same action whether google or themselves
           | processed it because they still get that actionable insight.
           | That actionable insight being actioned is what's detrimental
           | to you.
        
           | cscurmudgeon wrote:
           | Depending on the user's aims, Standard Oil's monopoly could
           | still be a net win.
           | 
           | - Some HN commenter in 1890s (probably)
        
         | somat wrote:
         | Certificate transparency has the same problem. It has the side
         | effect of sending every domain you visit to the certificate log
         | server. An advertisers wet dream. If you are using chrome this
         | is google. Now if you are using chrome I always assumed it was
         | sending everything you viewed to google anyway. So I guess this
         | is... OK, or at least, not worse than the status quo.
         | 
         | https://en.wikipedia.org/wiki/Certificate_Transparency
        
           | aaomidi wrote:
           | What?
           | 
           | > It has the side effect of sending every domain you visit to
           | the certificate log server.
           | 
           | No? It means that if a certificate is to be trusted, it needs
           | to enter a public append only log server.
           | 
           | It does not "send every domain you visit to CT." The
           | certificate simply has to be in CT before chrome trusts it.
           | 
           | CT logs are public, you can (and many do) make archives of
           | it.
           | 
           | The idea is simple, if a website is to be _publicly trusted_,
           | it needs to also be publicly logged.
           | 
           | You have heavily misunderstood what CT does, and the paradigm
           | of control around it.
        
             | salawat wrote:
             | Uh huh. And if I log what addresses reach out to me to
             | check for the presence of a particular certificate in the
             | logs?
             | 
             | You're not thinking about it hard enough. Your CT's would
             | effectively have to be implemented in a way where everyone
             | had a locally cached copy of the log to check against. That
             | at least shuts down an browsing activity leak vector.
             | 
             | If you centralize it, the abuse is a matter of when, not
             | if.
        
               | aaomidi wrote:
               | > Uh huh. And if I log what addresses reach out to me to
               | check for the presence of a particular certificate in the
               | logs?
               | 
               | Again, misunderstanding of how it works.
               | 
               | Logs are checked using the SCT in the certificate. It's a
               | signed "promise" from the CT that the pre-certificate
               | will eventually be appended to the logs in a few hours
               | since submission.
               | 
               | The various root programs monitor to ensure that the
               | promise does in fact happen.
               | 
               | Your browser process does *not* reach out to the CT log
               | on its own.
        
             | somat wrote:
             | Yes but how does your browser trust it? it has to check the
             | cert log. who has the server for the cert log. for chrome
             | this is google. Did you reach a different conclusion from
             | the big table in the wikipeda page where it requires "1 SCT
             | from a google log" ?
        
               | aaomidi wrote:
               | > Did you reach a different conclusion from the big table
               | in the wikipeda page where it requires "1 SCT from a
               | google log" ?
               | 
               | I've worked at both Let's Encrypt and Google Trust
               | Services. CT is something I'm intimately familiar with.
               | 
               | > it has to check the cert log.
               | 
               | No, it does not.
               | 
               | > Yes but how does your browser trust it?
               | 
               | Your browser ships with a log list: https://source.chromi
               | um.org/chromium/chromium/src/+/main:com...
               | 
               | The public key of the log is included there. It's checked
               | to see if the SCT is signed with the correct public key.
        
               | charcircuit wrote:
               | Chrome hasn't required "1 SCT from a google log" since
               | April 15th 2022
        
             | charcircuit wrote:
             | To be charitable, he could be talking about the
             | implementation of SCT auditing in Chrome, but it has never
             | uploaded all the sites you visit.                   There
             | are three possible states:              No Safe Browsing
             | protections -> no SCT auditing         Default Safe
             | Browsing protections -> SCT auditing logic selects a small
             | proportion of TLS connections and performs a k-anonymous
             | lookup on an SCT. If that privacy-preserving SCT lookup
             | reveals that the SCT is not known to Google but should be,
             | the client uploads the certificate, SCTs, and hostname to
             | Google (but no other information).         Enhanced Safe
             | Browsing protections -> SCT auditing logic selects a small
             | proportion of TLS connections and uploads the certificate,
             | SCTs, and hostname to Google (but no other information).
             | 
             | https://groups.google.com/a/chromium.org/g/ct-
             | policy/c/Fddjj...
        
               | aaomidi wrote:
               | Got it. Just for reference for others, SCT Auditing is
               | different from SCT validation.
               | 
               | I was talking about the validation side. SCT Auditing
               | basically helps ensure that the logs are keeping their
               | promise by letting the root program (Chrome, and Apple)
               | see what certificates are flowing on the internet and if
               | the merkle tree structure is valid.
        
           | olliej wrote:
           | No it absolutely does not. I think you're confusing CT with
           | OCSP, which is the browser checking to see if a certificate
           | has been revoked, which died a death not just because it was
           | incredibly slow and unreliable (CA's OCSP server fell over
           | enough that OCSP failed open), but more critically because it
           | was an incredibly invasive privacy problem.
           | 
           | CT does not require that, that was part of the explicit
           | design goal of CT. To be trusted by a browser a CA has to
           | publish every certificate they issue to multiple CT logs.
           | When the certificate is published to a CT log, the log
           | provides a signed confirmation, and that is included in the
           | final certificate.
           | 
           | Every browser has a set of CT logs that they trust. When they
           | perform the trust evaluation on a certificate they verify the
           | certificate is correctly signed by the CA, and then they go
           | through the included CT annotations and find the annotations
           | from logs that the browser trusts and verify that those
           | annotations are signed with the respective log's public key.
           | 
           | At no point does the browser perform any other network
           | requests (again, a problem with OCSP is that the additional
           | network access during TLS handshakes was incredibly slow and
           | unreliable - if CT required this it would require _more_
           | fallible network loads). That makes CT faster, and resolves
           | the privacy hole OCSP forced.
        
         | abraham wrote:
         | > The Google-run proxy can observe the user's IP address but
         | not the websites being visited and the third-party proxy can
         | see the web servers being visited but not the IP address of the
         | visitor.
         | 
         | Google will likely already have the user's IP from sync, safe
         | browsing, etc, if Google's proxy doesn't know what sites are
         | visited it doesn't sound like this increases Google's access to
         | any user data.
        
           | diroussel wrote:
           | Indeed that was my point it doesn't hurt Google just their
           | competitors. Thus concentrating power and increasing their
           | monopoly grip on the web.
        
       | isodev wrote:
       | That's an excellent example of the conflict of interest of trying
       | to be an ad company and a proponent for privacy at the same time.
        
         | leoh wrote:
         | Worse, it's arguably anti-competitive behavior. Though, truth
         | be told, once I looked at the facts, I would prefer Google to
         | have my data siloed (which itself, yes is not good) than random
         | third party data collectors that would sell it on data markets.
        
         | intelVISA wrote:
         | There's no conflict of interest, it's simply denying the enemy
         | from pillaging big G's harvest (you).
        
       | sacrosanct wrote:
       | > The Movement for an Open Web (MOW), an organization that has
       | lobbied against Google's Privacy Sandbox initiative by claiming
       | it's harmful to rival internet advertising businesses
       | 
       | They have much more to be annoyed at. VPN companies, The Tor
       | Project, AD blockers, etc
       | 
       | I use the Google One VPN[0] to cloak my real IP, but only
       | sparingly. For most of my Internet surfing I use a two hop VPN
       | setup. One VPN router with kill switch mode, and then I connect
       | to another VPN service on top of that, a sort of fake Tor /
       | private relay setup.
       | 
       | I don't funnel all my traffic into Google One's VPN. I like to
       | compartment and not put all my eggs in one basket. Looks like
       | I'll be doing the same when this new Google-owned IP cloaking
       | feature ships.
       | 
       | [0] https://one.google.com/about/vpn
        
         | yjftsjthsd-h wrote:
         | >> The Movement for an Open Web (MOW), an organization that has
         | lobbied against Google's Privacy Sandbox initiative by claiming
         | it's harmful to rival internet advertising businesses
         | 
         | > They have much more to be annoyed at. VPN companies, The Tor
         | Project, AD blockers, etc
         | 
         | While I'm sure they dislike those, the key word here is _rival_
         | - the claim is that Google has its own ad business, and is only
         | deploying privacy features in a way that hurts other ad
         | companies.
        
         | PrimeMcFly wrote:
         | That seems unnecessary. Why not just forward through an
         | anonymous SSH box?
        
       | leoh wrote:
       | We really need to be thinking bigger here -- Google included, to
       | the degree there are pro-privacy folks in the org (which there) -
       | in coming up with something that protects a lot more than the
       | identity via IP of Chrome users, but all web traffic.
       | 
       | I mean, maybe even rethinking IP, TCP/IP entirely.
        
         | LinuxBender wrote:
         | In my opinion this would just be a cat and mouse arms race.
         | Make something new and corporations will race to be the first
         | to add their tracking to it.
        
         | Andrex wrote:
         | I always thought there was something that could be done with
         | the BitTorrent protocol... I don't know how well it would work,
         | but an Internet that isn't so brittle mirrors are commonplace
         | and accepted wisdom would be nice.
         | 
         | BT isn't anonymous but a TCP/IP replacement version could be.
        
       | muteor wrote:
       | The anti abuse points from their github do not sound convincing
       | to me. There will be a high value in farming accounts to either
       | spam or attack. What do you do when the google proxy is dosing
       | your service?
        
       | sdefresne wrote:
       | Isn't this similar to what Apple does with Safari on iPhone where
       | they can hide your ip address by using iCloud servers as relay?
       | 
       | Discussed here: https://news.ycombinator.com/item?id=31387019 or
       | https://news.ycombinator.com/item?id=27467798
       | 
       | Why is it good when Apple does it but terrible when it is Google?
        
         | Lio wrote:
         | Is it because Apple's motivation is perceived to be selling
         | protection to its hardware customers where as Google's primary
         | motivation is perceived to be to get a monopoly in the
         | surveillance business?
        
         | dilipdasilva wrote:
         | It is not ok for Apple or Google to do this while at the same
         | time operating an ad business.
         | 
         | If they feel this is in the best interest of the end user, then
         | they should divest of either their ad business or control of
         | the browser. Neither company is willing to do this. This IP
         | move is anticompetitive as it consolidates even more control of
         | the ad ecosystem in a handful of companies. Google's response
         | that they are placed at the same disadvantage as other third
         | parties is not accurate. Google controls the browser and so has
         | full control to communicate any data between the browser and
         | their servers, bypassing the proxies.
         | 
         | There is only one thing that drives these companies and that is
         | maximizing profits for the benefit of their investors. This
         | objective is fine. However, it is disingenuous for either of
         | these companies to hide behind the defense that they care about
         | the privacy of end users.
         | 
         | If Apple cared about the privacy rights of all humans, why do
         | they share all data belonging to their customers in China with
         | the Chinese government. The only reason is profits. Google also
         | shares all their customer's data with any government that asks.
         | 
         | If there were a thousand companies that each had access to a
         | tiny sliver of a consumers data, we would have a system that
         | naturally protects end user privacy. However, with a few
         | companies controlling the vast majority of the consumer tech
         | landscape, we now have a system where a few for-profit
         | companies are keepers of our data and already sell out when
         | their profits are at stake.
        
         | olliej wrote:
         | From other comments it _sounds_ like Google's system is done as
         | a single proxy, which is bizarre to me because it means google
         | can see every site that is loaded which even for google seems
         | on the nose.
         | 
         | Apple's service is explicitly designed to prevent this exact
         | problem. There's a write up for it on apple's security site
         | (possibly part of the system security doc?). There are
         | intentionally two layers, the connection from the device ->
         | apple's servers, and then the connection from apple's servers
         | to Akamai or cloud flare (or some other CDN). The connection to
         | apple's servers is encrypted to a key from the 2nd layer CDN so
         | apple can't read it, that request is forwarded to the CDN which
         | decrypts it makes the request, then encrypts the response to
         | the client's key and sends that to apple, apple forwards that
         | encrypted blob on to the originating device which can then
         | decrypt it.
         | 
         | The end result is apple cannot ever see the destination or
         | response, and the backend CDN can't see the device that made
         | the request. That should be the design of _any_ privacy
         | conscious proxy service (including all the questionable
         | "privacy!" VPNs). That's kind of why I'm surprised that the
         | claim is that Google's service is a single layer - it's so
         | blatantly invasive.
        
       | gumballindie wrote:
       | Welp Firefox it is then.
        
       ___________________________________________________________________
       (page generated 2023-11-12 23:01 UTC)