back to article Google's Chrome is about to get rather in-your-face about HTTPS

Google and Firefox have been key drivers in the quest to get more people using HTTPS online, and starting this week the hammer is coming down. In a speech at Usenix Enigma 2017, Emily Schechter, a product manager for Chrome security, said that progress on HTTPS adoption was going well – currently over half of the top 100 …

  1. Gene Cash Silver badge

    Google too much in my face

    Google has also made Chrome DRM (Widevine) mandatory, so they can go rot along with Firefox. I endured Mozilla screwing with the UI for years, until they removed the ability in 44 to manage per-site cookie permissions (aka "ask every time")

    http://boingboing.net/2017/01/30/google-quietly-makes-optiona.html

    Thanks to the other El Reg commenters for sending me to Pale Moon.

    1. Baldy50

      Re: Google too much in my face

      Don't they gather info too, like Google and send it back to China?

      In 2016 computer security researchers from Fidelis Cybersecurity and Exatel discovered the browser surreptitiously sending sensitive browsing and system data—such as ad blocker status, websites visited, searches conducted and applications installed with their version numbers—to remote servers located in Beijing, China. According to Maxthon, the data is sent as part of the company's 'User Experience Improvement Program' and that it is "voluntary and totally anonymous." However, researchers found the data still being collected and transmitted to remote servers even after users explicitly opted-out of the program. The researchers further found the data being transmitted over an unencrypted connection (HTTP), leaving users vulnerable to man-in-the-middle attacks. Fidelis' Chief Security Officer, Justin Harvey, noted the data "...contains almost everything you would want in conducting a reconnaissance operation to know exactly where to attack. Knowing the exact operating system and installed applications, and browsing habits it would be trivial to send a perfectly crafted spearphish to the victim or perhaps set up a watering hole attack on one of their most frequented websites. Wiki.

      I prefer Firefox with whatever changes they make to the rest.

      1. Anonymous Coward
        Anonymous Coward

        Re: Google too much in my face

        I prefer Firefox with whatever changes they make to the rest.

        I would indeed not touch anything from Google with a 10ft barge pole, but I hope that Mozilla stops the nannying in FF when it finds a cert it doesn't like. Sometimes you know what the problem is, or you are simply testing and they've made it stupidly hard for people who know what they're doing to accept the cert anyway (temporarily or permanent) and look at the site regardless.

        It's very, VERY irritating and basically renders it useless as an engineering and security tool.

        1. td0s

          Re: Google too much in my face

          What's hard about clicking advanced then proceed anyway?

    2. Anonymous Coward
      Anonymous Coward

      Re: in my ChromeV56.0.2924.76(64bit)

      in my fully updated Mac (El'Capitan10.11.6) chrome://plugins/, currently Widevine seems to be OFF

      (tho' there are two hyperlinked "ENABLE" buttons "chrome://plugins/#")

      Widevine Content Decryption Module - Version: 1.4.8.962 (Disabled)

      Enables Widevine licenses for playback of HTML audio/video content. (version: 1.4.8.962)

      Name: Widevine Content Decryption Module

      . . .

      Location: /Applications/Google Chrome.app/Contents/Versions/56.0.2924.76/Google Chrome Framework.framework/Libraries/WidevineCdm/_platform_specific/mac_x64/widevinecdmadapter.plugin

      and furthermore - the 'old' trick of replacing "~/widevinecdmadapter.plugin" with summat'else might work?

      or delving into "chrome://plugins/plugins.js", but then DCMA?

  2. Anonymous Coward
    Anonymous Coward

    Double agenda?

    I think they're seriously overdoing it. So now a website which doesn't use HTTPS gets labeled insecure by default? Even if that website doesn't even ask it's users for any credentials or such? That's plain out stupid. As to the safety of HTTPS itself, anyone already forgotten about all those rogue CA's which started releasing valid certificates for all sorts of domains?

    Speaking of which: why not push for the acceptance of self signed certificates? I mean, if I go to a website "website.org" which is using a certificate issued by 'website.org' then isn't it a tad obvious that we're dealing with the same party? I mean, it's only encryption which is the main issue here. And that can also be easily handled by self signed certificates.

    It's only those certificate vendors who try to generate more revenue for themselves which started all that nonsense identity hype. I'm sure we can do without that easily.

    1. Adam 1

      Re: Double agenda?

      It *is* insecure. Whether that matters to you or not is another thing, but a http link allows a MitM to:

      1. Read and manipulate any content the site sends to you, removing anything they don't like and adding any they want. This may be as simple as ad substitution or could directly implement an exploit.

      2. Read and manipulate any content that you submit to the site.

      What is wrong with letting people know?

      And your self signing signature idea doesn't have legs because I can create a self signed signature for website.org and then MitM you. A CA needs to validate you control the domain. For example, letsencrypt will request you to host a file in a certain location to prove that the domain is under your control.

      1. gnasher729 Silver badge

        Re: Double agenda?

        If it's http, you don't need a "man in the middle" for an attack. Just a man.

        1. This post has been deleted by its author

        2. 's water music

          Re: Double agenda?

          If it's http, you don't need a "man in the middle" for an attack. Just a man.

          When I was little my dad always told me that he needed a man's courage. Nothing else though

      2. Anonymous Coward
        Anonymous Coward

        Re: Double agenda?

        A CA needs to validate you control the domain. For example, letsencrypt will request you to host a file in a certain location to prove that the domain is under your control.

        True, but that imposes a responsibility on the CA that some are not trustworthy enough to handle. That would be like handing Donald Trump the Presidency. Oh, wait ..

      3. Anonymous Coward
        Anonymous Coward

        Re: Double agenda?

        Yet Let's Encrypt or the like won't hinder many hosted site to be easily p0wend and send you phishing and malware inside a beautiful TLS session. Actually, MitM a commercial network may be more difficult than p0wning vulnerable sites (unless, of course, you're a TLA).

        After all, all this TLS frenzy began after Snowden whistle blowing. It's far more aimed to ensure your data are not easily snooped by TLAs, than ensuring you get "clean" safe data from "clean" safe sources.

        If you fully trust whatever is served even inside a TLS connection, especially now getting a certificate it's easy and cheap, you're going to awake painfully, one day.

      4. Drew 11

        Re: Double agenda?

        Not if DANE is used, but Google and Firefox don't want to give you too much control over your own destiny.

      5. Anonymous Coward
        Anonymous Coward

        @Adam1

        "And your self signing signature idea doesn't have legs because I can create a self signed signature for website.org and then MitM you. A CA needs to validate you control the domain."

        You mean like those rogue CA's which will easily give you a signed certificate for existing domains like google.com? It's not as if HTTPS fully rules out any risk of a man in the middle attack as you make it sound.

        1. Adam 1

          Re: @Adam1

          I don't think what I'm proposing would be required to break your suggestion is at all beyond the skillsets of anyone who reads a tech news site.

          1. Buy the applicable hardware.

          (Eg https://wifipineapple.com/ )

          2. Create a self signed certificate for website.org

          (Eg https://technet.microsoft.com/en-us/library/cc753127(v=ws.10).aspx )

          3. Look for a location likely to have free wi-fi but who fail to use HTTPS

          (Eg https://t.co/6Bu4v9f5Qn )

          4. Redirect any form submit action to a Uri under your control

          (I won't detail that step RTFM)

          5. Profit

          Alternatively, pick a café/library/train station/hotel and call you fake AP "Free McDonald's WiFi", hijack the first HTTP page they request, put the McDonald's logo on top and say "Sign in with Facebook", put the f logo on it and many people will just connect to it and type in their credentials.

          CAs are imperfect. Diginotar and Wosign stand out, but I couldn't characterise them giving me a fake cert as "easy". Having the right political connections to get them to make a fake cert for you is much less of a threat for most people than what I have described above.

          CAs are like democracy. The worst form of government, except for all the other forms we have tried from time to time.

    2. dajames

      Re: Double agenda?

      ... if I go to a website "website.org" which is using a certificate issued by 'website.org' then isn't it a tad obvious that we're dealing with the same party?

      Obvious, maybe, in a misleading kind of a way.

      If someone has hacked DNS so that you are directed to a server that isn't "website.org" but has a self-signed certificate claiming that it is, then you tend to accept that even though it's a lie. If they have to have a certificate signed by a well-known CA the subterfuge becomes apparent.

    3. DaLo

      Re: Double agenda?

      Certificates are for encryption and authentication.

      Self-sign certificates just provide a level of encryption. They are not secure on a public internet and provide no authentication.

      Why would you need them anyway, wildcard certificates are cheap and free certificates are available.

      The trusted root does have to be trusted for sure, but you can revoke trusted roots yourself, however if they screw up then they can lose their whole business if they are deselected by top browsers.

      Saying a site is 'not secure' is just stating a fact. So if it asks you to send any data in a form field or it is hosting content that you may not wish to be associated with the it is a reminder. As more sites are going TLS then you start to accept it and gloss over unsecure sites.

      Just remember that an unsecure site will also be receiving a lot of data about you that could be intercepted or could have any amount of malicious code injected into it.

    4. DaLo

      Re: Double agenda?

      At the moment the current build (56) of Chrome will only mark your site as not secure if you go to a page which asks for a password (input type-password) or it detects a field asking for a credit card number and you don't have https on either the parent or any sub frames (including iframes). The article is a bit misleading.

      It is slated that it may implement it at some time n the future (I heard October) for all sites on all pages if not secure.

      To see and example of the browser bar warning if using chrome look here (http://http-password.badssl.com/)

    5. Marco Fontani

      Re: Double agenda?

      So now a website which doesn't use HTTPS gets labeled insecure by default?

      Only if it contains something that looks like a log-in or sign-up form, for now

      … as a man-in-the-middle would easily be able to change where the insecure form goes to, and you'd be none the wiser you'd be sending your personal details or log-in details to Mallory

  3. Nick Kew

    Cult of useless HTTPS

    HTTPS offers some security, but with dangerous points of failure (like CAs) that should be familiar at least to regular readers.

    However, statements like this indicate a bandwagon onto which people are thoughtlessly jumping:

    "Incidentally, The Register can be viewed over HTTPS, from our forums login to white papers to editorial articles – hats off to our tech team for that."

    The cost of that, in terms of loss of cacheability, is akin to that of a stoppage on the trains driving millions of commuters into cars. Why is a site whose contents are public imposing that cost on the 'net?

    Even forums has nothing more than low-value passwords to protect: if someone impersonates me here, they haven't got anything like access to anything potentially valuable like my money or private communications.

    1. Anonymous Coward
      Anonymous Coward

      Re: Cult of useless HTTPS

      Without it shitty consumer grade ISPs will ALWAYS be tempted to inject adverts into pages.

      Even for totally public webpages you want to ensure the integrity of the data, and know who to blame for malvertising (aka advertising).

      1. inmypjs Silver badge

        Re: Cult of useless HTTPS

        "Without it shitty consumer grade ISPs will ALWAYS be tempted to inject adverts into pages."

        And not so shitty IPSs might offer to outject adverts for you, especially on mobile connections.

        That is why google are pushing this so hard- they don't want anyone messing with or removing their adverts.

    2. Adam 1

      Re: Cult of useless HTTPS

      Users are users and will reuse those passwords on other sites. I agree on the caching problem. That is a solvable problem if they hash all the resources used by the page then sign the hash with their private key but i guess noone is pushing for it.

    3. rh587

      Re: Cult of useless HTTPS

      Caching is a problem, but users will reuse passwords and shitty ISPs will inject ads. You're protecting them from themselves and from their ISPs.

      Additionally, and perhaps most importantly for a business - Google are adding https to PageRank, meaning there is a commercial incentive to operate https if you want to be found.

    4. DaLo

      Re: Cult of useless HTTPS

      "The cost of that, in terms of loss of cacheability, is akin to that of a stoppage on the trains driving millions of commuters into cars. Why is a site whose contents are public imposing that cost on the 'net?"

      "I agree on the caching problem."

      "Caching is a problem"

      What caching problem? HTTPS pages and resources are cached like normal.

      1. Tom 38

        Re: Cult of useless HTTPS

        What caching problem? HTTPS pages and resources are cached like normal.

        You are thinking of browser caches. HTTPS makes edge caches trickier.

        1. DaLo

          Re: Cult of useless HTTPS

          How would edge caching be equivalent to "... that of a stoppage on the trains driving millions of commuters into cars."

          I have seen very few using a proxy with caching where they don't also use DPI and company trusted certs anyway.

        2. Adam 52 Silver badge

          Re: Cult of useless HTTPS

          Why does it make caching hard? You terminate TLS in the CDN and run a new SSL session from the CDN to the source. As a bonus you can validate the CDN cert on the source as well.

          [Sorry for the TLAs, but this is a tech sites]

          1. Tom 38

            Re: Cult of useless HTTPS

            Why does it make caching hard? You terminate TLS in the CDN and run a new SSL session from the CDN to the source. As a bonus you can validate the CDN cert on the source as well.

            Well, yes. Doing those things is harder than doing nothing, qed caching is trickier.

    5. a_yank_lurker

      Re: Cult of useless HTTPS

      The problem Chocolate Factory is trying to address is overall poor security on the web combined with the generally poor user skills. HTTPS is not perfect, no one with a clue will say that, but it does offer more security than nothing (HTTP). To some extent the idea is to slow down the miscreants. Also, another part of this is add more layers to get through - defence in depth.

      Remember the average reader of El Reg is likely very knowledgeable about computers, web design, etc. while average user only knows how to use a computer and is otherwise clueless about how they work.

  4. localzuk Silver badge

    Annoying

    For schools this is annoying. It increases panic in staff and students, and then ends up making them ignore all warnings about security.

    It also means we have to implement larger filtering solutions, as we end up having to intercept and scan more secure websites to ensure we are complying with our legal obligations.

    1. Adam 52 Silver badge

      Re: Annoying

      "as we end up having to intercept and scan more secure websites"

      Or insecure websites, as you've just made them. You are a man-in-the-middle so stop complaining about being outed as one.

      " to ensure we are complying with our legal obligations."

      What legal obligations would they be then? Can't think of anything that says children have no human rights, whatever headmasters would like to portray.

  5. Simon Brady
    Unhappy

    Security and obscurity

    So if the Chrome team's mission is to help users be secure, why has Chrome ca. 56 made it so much harder to view certificate details? Up until recently you could right-click on the "Secure" marker in the address bar and go straight to the cert - now all that gives you is a link to a generic help page, and you have to drill down into the Developer Tools UI to find this information.

    In what world is this an improvement?

    1. Boothy

      Re: Security and obscurity

      Judging by this chromium bug discussion.

      It seems the previous 'Details' button, which took you directly to the Security tab in the Dev tools window, was always just a temporary thing to bring attention to where the security info is located (i.e. inside the Dev tools).

      Looks like the plan was always to remove it after a while, as they seem to think regular uses don't need to view the certificates. (Not sure I agree with this myself!)

      The new 'Learn more' button is for regular users, rather than developers, so just takes you to the generic web page documenting what the various icons mean etc.

      Seems the chromium devs didn't like the idea of regular users landing in the dev tools: Quote '...so that regular users who are newly clicking on the lock icon (due to icon/verbose UI changes) don't click on it and end up somewhere unexpected.'

      Like I said, not sure I agree with this approach, and a few people on the chromium forum even suggest having a new link/button to take you directly to the certificate details.

      But for now, seems we are stuck with Ctrl+Shift+I, or the More Tools -> Developer tools menu, then click the Security tab :-/

  6. adam 40 Silver badge

    Browser Facism gone mad

    I'm sick of this continual upgrade madness.

    Chrome now won't even let me view sites with RC4 TLS, but will let me view plain http - where's the logic in that?

    Grrrr...

    1. Rich 11

      Re: Browser Facism gone mad

      The logic is that one is something which people might think is secure but which isn't necessarily so, while the other is something no-one (quite correctly) thinks is secure.

      1. Anonymous Coward
        Anonymous Coward

        Re: Browser Facism gone mad

        So flag them as "utterly unsecure" and require me special and complex ways to access them - but let me access them.

        I have some devices web management tools which were never upgrade beyond SSL3 and RC4. To use them I have only the option to remove any kind of encryption (sending auth data in the clear), or keep some old browser version around.

        The issue with Google and Mozilla morons is they can't see HTTP and HTML are unluckily used in too many situations, and not only on the InterWeb itself to serve ads to clueless users. While Google may replace all of its network devices and servers every year, unluckily we can't.

        1. SImon Hobson Bronze badge

          Re: Browser Facism gone mad

          I'm with you there - managing (for example) networking devices which as far as the manufacturer is concerned ceased to exist many years ago is a right PITA when various levels of software just won't let you. Java and the cretins responsible gets a special position in the hell I'd send some people to for not including a "yes I know", "yes I really know", "now FFS let me do my job" option to actually access these devices.

          1. Boothy

            Re: Browser Facism gone mad

            Personally, I think disabling things like obsolete security options by default in a mainstream browser like Chrome is a good idea, at least for the vast majority of users.

            Although I know what it's like trying to do development work with Chrome on older environments, as I've had issues myself with access etc.

            It would be nice if Chrome had a Developer/Super User type mode, where you could then re-enable things, perhaps on a case-by-case basis on a chrome://somename page.

            For security, you could even have it based on zones, i.e. specific addresses listed in zone 'XYZ' can use RC4 with TLS (or whatever is needed), anything not listed in zone 'XYZ' (or some other explicit zone) is automatically in the default zone, which still uses the default security, so leaves your Internet and other sites still locked down.

  7. tr1ck5t3r
    Trollface

    Hiding the delivery of malware to devices just gets easier and easier. MITM to spot your problems?

  8. Anonymous Coward
    Anonymous Coward

    Google keeps visability hides others

    Useful non identifiable analytics data from the isp's will become unavailable but search to site analytics will still be kept under google's lock and key! Is it all about security or control of competitive data?

  9. G2
    Facepalm

    old article published with today's date?

    when was this article written? it's a bit behind the news...

    quote:

    Later this week Chrome 56 will be released, and [...]

    /quote

    newsflash: Chrome 56 has been released almost a week ago, on january 25th... on my system it upgraded itself on Friday, ‎27 ‎january 2017 at ‏‎08:09:26 (timestamp of the 56.0.2924.76 folder)

    https://chromereleases.googleblog.com/2017/01/stable-channel-update-for-desktop.html

    1. Anonymous Coward
      Anonymous Coward

      Re: old article published with today's date?

      We'll ask it's from a conference that run's from the 30th of Jan to 1st of Feb, not that old. (there is a very helpful link in the article those shows the dates in f'ing big letters).

      https://www.usenix.org/conference/enigma2017

      A simple typo and it should be V57?

  10. Anonymous Coward
    Anonymous Coward

    Will the authors/editors please stop propagating this outright lie that "SSL is free, or very cheap". Yes you might be able to get cheap (I use SSLMate) or free certificates IF YOU DO YOUR OWN HOSTING. But for the majority of us using third party hosting, SSL is either not an option, or we have to pay whatever the ISP charges.

    One ISP we use at work (hence posting as AC) charges around £50/year for an SSL cert on your site. For a provider we use, we've just had to pay them around £100 to put a new cert on the portal for a year.

    Might be free or cheap for the likes of Google - but back in the real world it generally is anything but.

  11. Displacement Activity
    Thumb Down

    Follow the money

    1 - Google charges for TLS on inbound connections;

    2 - Google is a prime mover behind 'TLS Everywhere', and is now starting to factor this into page rankings;

    3 - (Google's) Let's Encrypt certificates prove exactly nothing except that you have control of the server for which the certificate was granted (you only have to post stuff on it to get the certificate);

    4 - Bad People control their own servers anyway, so can trivially get their own certificates; MITM is therefore irrelevant on these sites

    95% of sites have exactly *no* reason to worry about whether someone is forging their site, or whether there's a MITM somewhere in the connection. So, Google is screwing us, and we have to pay the price by dicking about with TLS on our own sites, and keeping certificates up-to-date, and trying to ignore pointless warnings, and handing cash to them if we're stupid enough to host with them.

  12. Wolfclaw

    Chrome 55 spends so much time crashing on my Android, that I ended up with Opera Mini, it just works.

  13. Oengus
    Stop

    Don't do it

    Geolocation can reveal an internet user’s home or work address, and be used for tracking. Such data needs to be more secure, she argued, and that need is only growing stronger.

    So stop requesting the sensitive data unless truly appropriate, openly visible and approved by the end user. For the vast majority of websites there is no need for this sensitive information. Google et al only want it to be able to profile you and track you to their benefit (not yours).

  14. This post has been deleted by its author

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like