Slashdot Mirror

← Back to Stories (view on slashdot.org)

Google Will Give a Search Edge To Websites That Use Encryption

Posted by timothy on from the bit-of-an-incentive dept.

As TechCrunch reports, Google will begin using website encryption, or HTTPS, as a ranking signal – a move which should prompt website developers who have dragged their heels on increased security measures, or who debated whether their website was “important” enough to require encryption, to make a change. Initially, HTTPS will only be a lightweight signal, affecting fewer than 1% of global queries, says Google. ... Over time, however, encryption’s effect on search ranking [may] strengthen, as the company places more importance on website security. ... While HTTPS and site encryption have been a best practice in the security community for years, the revelation that the NSA has been tapping the cables, so to speak, to mine user information directly has prompted many technology companies to consider increasing their own security measures, too. Yahoo, for example, also announced in November its plans to encrypt its data center traffic.

33 of 148 comments (clear)

  1. Great step! by satuon · · Score: 5, Interesting

    That's a really great step from Google, I had never thought that it can be done in such a neat way. What's next? Can they also do it for IPv6?

    1. Re:Great step! by satuon · · Score: 2

      It could create an incentive to switch to a different ISP that supports it (where possible), which could in turn create an incentive for ISPs themselves to switch to IPv6.

    2. Re: Great step! by Nexus+Unplugged · · Score: 3, Informative

      CloudFlare has also announced that they're planning to roll out free SSL to customers in the coming months.

    3. Re:Great step! by defaria · · Score: 2

      Add to this that sometimes the reputable companies still don't do the right thing. As a large financial institution you'd think that Wells Fargo of all companies would at least be sporting an extended validation certificate! But they don't. And that's not their only fumbling in security (http://defaria.com/WF). I have my reasons for still dealing with them but I watch them like a hawk!

    4. Re: Great step! by petermgreen · · Score: 4, Informative

      They do BUT

      1: their rules on who can get the free certs seem to be varied and arbitary. I've seen reports of an opensource developer being given a free cert initially but then come renewal time told that merely having a donation button makes their site count as "ecommerce" and therefore ineligable
      2: they make the expiry artifically short (the CA industry as a whole does this but startSSLs free certs are epecially bad),
      3: they refuse to renew certs until just before they expire and refuse to reissue certs without revoking the old one.
      4: each free cert only covers a domain and one hostname under that domain (e.g. bar.com and foo.bar.com). This effectively means you end up needing one IP per hostname you want SSL on (until IE on XP becomes insignificant anyway).

      It's nice that there is a free (as in beer) option for some people but it's also clearly got a number of artificial restrictions on it to push people towards their paid options.

      --
      note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register
    5. Re: Great step! by Rich0 · · Score: 4, Informative

      StartSSL still give out free certificates to individuals right?

      Yes, as long as you don't change your certificate after the key is lost as a result of HeartBleed. If you want your users to be secure, then you need to pony up $25. How that isn't a violation of the Mozilla policies is beyond me. I can give StartSSL clear proof that a private key has been disclosed, and they won't revoke it unless somebody pays them to do it.

    6. Re: Great step! by Darinbob · · Score: 4, Insightful

      It's already monetary discrimination, since well design sights with interesting products will show up higher in the rankings than the local mom&pop web site where they could only afford to hire a high schooler to do the design.

      The whole point of ranking is not to make sure everybody is perfectly equal, but to help the customer find the most relevant results. If I'm searching for a bank then I most certainly want a bank with security to be ranked higher than one without. However, I can see the issue that it's only Google who gets to decide what's relevant. Perhaps there should be some user specified criteria, such as letting me decide to show only IPv6 capable sites.

    7. Re: Great step! by heypete · · Score: 3, Informative

      2: they make the expiry artifically short (the CA industry as a whole does this but startSSLs free certs are epecially bad),

      A validity time of one year is pretty standard for SSL certs (paid certs often charge per year). Could they issue them for 20 years? Sure, but a one year validity is not unusual. Class 2 certs are good for two years.

      3: they refuse to renew certs until just before they expire and refuse to reissue certs without revoking the old one.

      I get renewal notices two weeks prior to expiration. That's pretty reasonable. If I recall correctly, I can generate a new cert for my site any time in that two-week period, so I don't need to wait for the cert to expire before replacing it.

      While I wish they allowed free reissuance of certs at any time, I don't really see why requiring revocation is a showstopper.

      4: each free cert only covers a domain and one hostname under that domain (e.g. bar.com and foo.bar.com). This effectively means you end up needing one IP per hostname you want SSL on (until IE on XP becomes insignificant anyway).

      That's also the case for pretty much any of the inexpensive paid certs too. You can always get a wildcard cert but most CAs charge at least $100/year for a single wildcard cert. StartSSL charges $60 for Class 2 validation, and you can issue unlimited certs (wildcard or not). Organizations can get Class 2 certified for $120 ($60 for identity verification, $60 for organization verification) and can issue unlimited certs. For a company needing more than one cert, StartSSL is still cheaper.

      It's nice that there is a free (as in beer) option for some people but it's also clearly got a number of artificial restrictions on it to push people towards their paid options.

      Considering their paid certs are often cheaper than comparable offerings from other CAs, it doesn't really seem unreasonable to me. Doubly so because they're run by competent people who respond promptly to inquiries, even from free users. I've been a StartSSL customer for years (and also used other CAs like GoDaddy, Comodo, Thawte, etc.) and the customer service from StartSSL has always been excellent.

      If you don't want to get a StartSSL cert or they don't meet your needs, that's fine. NameCheap and others sell single-domain Comodo certs for $9/year. RapidSSL certs are a buck or two more per year. That costs less than a single beer at the local bar. Hardly a massive expense.

    8. Re: Great step! by swillden · · Score: 2

      However, I can see the issue that it's only Google who gets to decide what's relevant.

      Google gets to decide what's relevant in the rankings on their site, but not what's relevant for other search engines. If they do a bad job of picking good ranking criteria, it gives other engines an opportunity to provide better service. This is a somewhat coarse mechanism for demanding more relevant criteria, I suppose, but you'd better believe that Google takes it very seriously. They have a lot of other signals that help them decide whether users are well-served by the top-ranked hits, and if something like preferring HTTPS damages that, it'll almost certainly lose.

      --
      Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
  2. So now Google establishes Internet standards by neilo_1701D · · Score: 4, Insightful

    I'm not convinced that this is a good precedent. Sure, they're encouraging sites to use HTTPS today... but what about tomorrow?

    Speculation: Websites that block competing search engines from indexing their content may rank higher in Google searches? Websites that process payments using Google rank higher in Google search?

    I'm not saying that HTTPS is a bad thing... but once they open the door once to arbitrary ranking changes done on a whim, that door can be opened again.

    1. Re:So now Google establishes Internet standards by Anonymous Coward · · Score: 3, Insightful

      As opposed to the currently non-arbitrary ranking algorithm? What the hell are you talking about.

    2. Re:So now Google establishes Internet standards by Anonymous Coward · · Score: 2, Insightful

      Don't like it, use Bing!

    3. Re:So now Google establishes Internet standards by Agent+ME · · Score: 5, Funny

      They've already been using their ranking system to encourage HTTP and HTML. Think of all the poor BBSs and gopher servers they've been discriminating against!

    4. Re:So now Google establishes Internet standards by bill_mcgonigle · · Score: 3, Interesting

      Google has been using dozens of quality metrics for years to adjust its rankings. This isn't a new concept.

      It's not clear to me which HTTPS configurations it's favoring, though. Is Strict Transport Security a requirement? People with high-longevity system needs are going to need to upgrade to EL7 to make good HTTPS feasible, so there will be a transition period.

      As far as standards - look, W3C, IETF, et. al. have completely failed to keep up. From 1993 to 1997 we went from HTTP 0.9 to to HTTP 1.1, which is where we are today. HTTP 2.0 will have been languishing for two decades by time there's a standard and any significant adoption. That's not Internet-time.

      Google has made some mistakes with SPDY and QIC but at least they're actually trying to move the ball down the field instead of just arguing on the sidelines. It used to be that lots of players would do the same thing and fairly quickly a concensus would emerge. We have a serious breakage problem in the current community process. Google is doing it right - it's everybody else that's not.

      --
      My God, it's Full of Source!
      OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
    5. Re:So now Google establishes Internet standards by satuon · · Score: 2

      I wish they would index real FTP servers, not just those with an HTTP interface.

    6. Re:So now Google establishes Internet standards by Anonymous+Brave+Guy · · Score: 3, Insightful

      While your points about the snail's pace of web "standards" development are fair, it's also important not to go too far the other way. Not so long ago, another browser became dominant in market share through pushing new but not widely supported features its own way, and people started making web sites that were written specifically to work with that browser rather than any common standard.

      That browser was Internet Explorer in the late 1990s, and the result was IE6.

      --
      If you disagree, post your argument. (-1, Overrated) isn't your personal censorship tool for views you don't like.
    7. Re:So now Google establishes Internet standards by WaffleMonster · · Score: 2

      As far as standards - look, W3C, IETF, et. al. have completely failed to keep up. From 1993 to 1997 we went from HTTP 0.9 to to HTTP 1.1, which is where we are today.

      Most HTTP 1.1 features are useless. If it disappeared tomorrow nobody would care or even be able to tell it has gone missing.

      HTTP 2.0 will have been languishing for two decades by time there's a standard and any significant adoption. That's not Internet-time.

      The pace of standards development is driven by commercial need rather than abstract notions of staleness, "the future", "progress"..etc.

      The only reason for delay is nobody cares. The incremental benefit is so trivial as to not be worth the effort unless you happen to be Google. When people care shit gets done even if it means draft implementations making their way into production.

      Google has made some mistakes with SPDY and QIC but at least they're actually trying to move the ball down the field instead of just arguing on the sidelines.

      My personal opinion we are much better off working TCP and TLS extensions to reduce round trip delays. You can for example in best case get a secure HTTPS request to server without completing a single round trip leveraging TCP and SSL features (fast open, session tickets) neither of which requires maintaining server state, as would keeping TCP sessions open longer than absolutely necessary or
      having to suffer HOL penalties or get weighed down by pointless politics and scope creep (opportunistic encryption)

      Finally working transport and security layers has added benefit of being instantly useful to all protocols not just TCP.

      We have a serious breakage problem in the current community process. Google is doing it right - it's everybody else that's not.

      The "community" is like the UN. It is simply a forum for those with power (e.g. commercial interest) to negotiate... nothing more nothing less.

    8. Re:So now Google establishes Internet standards by just_another_sean · · Score: 4, Insightful

      but once they open the door once to arbitrary ranking changes done on a whim, that door can be opened again.

      Was that door ever closed? They're ranking algorithm has been arbitrary since the beginning and has changed very frequently over the years in an effort to reduce gaming the system and to generally improve results. If anything I'd say it's nice that they're at least telling people about this change vs. just quietly adjusting things and leaving site owners to wonder what happened to their page rank.

      --
      Creationist Textbook Stickers Declared Unconstitutional by CowboyNeal
  3. on advice of counsel.... by turkeydance · · Score: 2

    i, Google, (corporations are legal individuals in USA) refuse to rank my response due to it's incriminating nature.

  4. It's about time! by mcrbids · · Score: 4, Interesting

    Expensive advertising campaigns engender trust because it shows that the advertiser has the resources to carry out the campaign. It's why online ads are so commonly ignored - people want to do business with "reputable" companies and expensive advertising is a way of establishing repute.

    Similarly, putting out the modicum of effort to perform basic security like SSL is a signal that the website is reputable. I mean, if you can't be bothered to buy a $50 SSL certificate and install it, are you *really* trustworthy?

    SSL should be a basic signal of trustworthiness.

    --
    I have no problem with your religion until you decide it's reason to deprive others of the truth.
    1. Re:It's about time! by WaffleMonster · · Score: 3, Interesting

      Similarly, putting out the modicum of effort to perform basic security like SSL is a signal that the website is reputable. I mean, if you can't be bothered to buy a $50 SSL certificate and install it, are you *really* trustworthy?

      LOL and here I thought all this time the Internet was supposed to reduce costs and barriers to competition... yet here we go "the higher the fewer".

      When your making the big bucks off Google by operating industrial scale link farms $50/year is a small price to pay for success.

      Someone please remind me again why we are even contemplating enriching the clusterfuck that is the CA industry which sees no problem with use of completely automated systems and non-existent documentation requirements prior to issuing certificates?

  5. Thanks to Google and the NSA ! by Anonymous Coward · · Score: 2, Interesting

    Thanks to Google for making the web a little bit more secure by promoting secure websites!
    Thanks to the NSA for tapping the web so blindly and boldly than we should react!

    If the NSA was not so bold and had tapped only these who were under suspicion of bad behavior, the status-quo would have been kept. Now the privacy of everyone is a little bit more secure and the NSA will have a little bit harder times managing MITM attacks on every netizens.

    An EU Citizen who like its privacy.

  6. Cat blog by ZipK · · Score: 3, Insightful

    So my cat picture blog will rank lower than a competitor's SSL encrypted cat picture blog, even though neither of us require you to log in or even prove you are a cat?

    1. Re:Cat blog by Cyberdyne · · Score: 4, Informative

      Still, HTTPS would at least prevent your ISP from monitoring your browsing activity.

      That's part of it - a valuable enough part in itself, IMO; at least one UK ISP, TalkTalk, has started duplicating HTTP requests made by their customers: so, if you request http://example.com/stuff on one of their lines, 30 seconds later they'll go and request the same URL themselves for monitoring purposes. Obviously, enabling SSL prevents this kind of gratuitous stupidity - and the previous incarnation of such snooping, Phorm. If enough websites enable SSL, ISPs will no longer have the ability to monitor customer behavior that closely, all they will see are SSL flows to and from IP addresses, and whatever DNS queries you make to their servers, if any. (Use encrypted connections to OpenDNS or similar, and your ISP will only ever see IP addresses and traffic volume - exactly as it should be IMO!)

    2. Re:Cat blog by IamTheRealMike · · Score: 5, Informative

      Yes, for news and such it doesn't make that much sense. Still, HTTPS would at least prevent your ISP from monitoring your browsing activity.

      It's actually a lot more than that. HTTPS isn't just about protecting passwords anymore, not post Snowden.

      Let us recall one of the more interesting things we learned about SSL via the NSA leaks: the Five Eyes countries apparently have not broken SSL yet despite that the internet is still not capable of stopping them. The reason is a system they've built called QUANTUM.

      QUANTUM is a series of systems that work together. Imagine it like being a giant set of guard towers on the internet backbone. QUANTUM is called that because it's based on deep packet inspection and insertion. The first part is a massive set of DPI devices that trawl unencrypted internet traffic passing through intercept points. These DPI devices can be configured by NSA/GCHQ analysts to look for selectors - personal identifiers like email addresses, IP addresses, cookies and so on. QUANTUM does not run on every internet link and cannot see through encrypted traffic, but that doesn't matter: it's like a searchlight crawling the grounds of a prison at night. It doesn't matter that it can't light up everywhere simultaneously - once tasked it will keep searching until it finds you. Given enough time and good selectors, it will always find you, simply because the average internet user makes many different unencrypted connections to many different websites.

      Once QUANTUM locates an un-SSLd traffic stream that matches your selectors, the next step begins, this is called QUANTUM INSERT. You see these DPI devices are not only capable of reading traffic but also injecting packets directly onto the backbone as well. This allows them to race legitimate answers from the real servers, and redirect the victim to an entirely different server (this is probably based on racing DNS lookups although I think the leaked docs were fuzzy on this aspect). These races are called "shots" and interestingly, they don't always succeed - sometimes the NSA is slower than the real server. But QUANTUM keeps trying and eventually you end up connected to this new FOXACID server, which then proceeds to act as an HTTP proxy for the real request and injects an exploit kit. That then pwns your system such that the NSA can now see all your encrypted traffic, along with turning on your microphone and so on.

      An observant reader will notice something very important about the above description. The longer you can stay in the SSLd web, the longer it will take for QUANTUM to hack you. That means you directly benefit from a website being SSLd even if all it contains is cat pictures and you don't even log in. Once QUANTUM has figured out your IP address, any non-SSLd HTTP connection is a useful foothold.

  7. OK fine but give us a free CA by Cthefuture · · Score: 5, Insightful

    I have no technical problem switching every website/server I have to SSL but the actual problem is the price of all those SSL certs. Most of my sites are just hobby type sites that I run for my own enjoyment and to benefit others (quite a few "others" I should mention; some of my sites are very popular). However, I don't make any money off these, in fact it already costs me money to run them.

    Now you want me to add SSL so that people can still find my relevant and useful information? Well, OK but how the hell am I suppose to pay for it? SSL server certs are expensive. The whole thing is a scam to make the few "official" CA's rich. How about some sort of official public service that can hand out server certs of every registered domain? Every domain should come with an unlimited supply of SSL certs or at least a wildcard cert and a renewal service, free of charge.

    --
    The ratio of people to cake is too big
    1. Re:OK fine but give us a free CA by RobinH · · Score: 3, Insightful

      Agreed, if Google wants to do this, maybe they should also become a free Certificate Authority. Wouldn't that tick off the Verisigns of the world...

      --
      "I have never let my schooling interfere with my education." - Mark Twain
  8. Android Browser 2.x and IE/XP lack SNI by tepples · · Score: 4, Informative

    I mean, if you can't be bothered to buy a $50 SSL certificate and install it, are you *really* trustworthy?

    It's not only the cost of a certificate, which StartSSL provides without charge to individuals. It's also a dedicated IPv4 address if you want to reach people still using Android 2 or Windows XP. A lot of entry-level hosting packages use name-based virtual hosting, and doing this over name-based virtual hosting requires the TLS stack to support Server Name Indication (SNI). Android Browser didn't gain support for SNI until Honeycomb (3.x) on tablets and ICS (4.0) on phones, and Internet Explorer didn't gain support for SNI until Windows Vista.

  9. StartSSL or DANE by tepples · · Score: 2, Informative

    How about some sort of official public service that can hand out server certs of every registered domain?

    You mean like StartSSL? Or what about DANE, which stores TLS certificates in DNSSEC?

    1. Re:StartSSL or DANE by petermgreen · · Score: 2

      You mean like StartSSL?

      Hardly an official service, just a commerical CA that hands out freebies to some but not all sites that ask for them and puts technical restritions on those freebies which push people to either buy the commercial products or spend more on hosting (do I pay for n extra n IPv4 addresses or do I pay for a wildcard cert).

      Or what about DANE [wikipedia.org], which stores TLS certificates in DNSSEC?

      Sadly not implemented anywhere near widely enough to be useful.

      --
      note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register
  10. HTTP-only ad networks by tepples · · Score: 3, Informative

    Slashdot makes HTTPS available only to subscribers because historically, web ad networks haven't supported HTTPS. Only in September 2013 did Google AdSense roll out HTTPS support.

  11. Server Name Indication by tepples · · Score: 2

    SSL requires a dedicated IP

    Only if your clients include Android 2.x or Internet Explorer on Windows XP. Every other browser that matters supports Server Name Indication (SNI), which allows name-based virtual hosting to work through TLS. As of today, if you can see my site without certificate errors, your browser supports SNI.

    and an extra charge

    StartSSL issues certificates to individuals without charge.

    Is there really a privacy concern if my visit to a weather site, a dictionary, or other factual content site is not encrypted?

    Yes. Someone could copy and replay the session ID linked to your user account on the site and gain your privileges.

    Then there's the bandwidth issue. Sites that go SSlL will use more bandwidth

    What in TLS introduces this substantial extra overhead? And how much overhead is it, really? I do know of a common misconception that HTTPS isn't cacheable. In fact, a document delivered through HTTPS is cached on the client the same way anything else is cached on the client. It just isn't cached on an intermediate transparent proxy, which hurts if your ISP is using such a proxy to cut down on its own upstream.

  12. Re:FTP authentication by satuon · · Score: 2

    Anonymous login accepts any password, just put a random string. As for the LIST command, if FileZilla can read it, so can Google, they're not morons. You just handle all the possible variations of all the popular FTP servers. Yes, you actually have to write some code, but last I heard Google has programmers on staff.