9 ms·
Facebook eventually addressed the issue by making the site accessible over HTTPS—though, as the authors of the 2008 paper note, HTTPS can be a "rigid and costly
by nkurz 14y ago
Facebook eventually addressed the issue by making the site accessible over HTTPS—though, as the authors of the 2008 paper note, HTTPS can be a "rigid and costly" solution.
This same excuse has existed for about as long as HTTPS, which dates to Netscape Navigator 1. Is it still that "rigid and costly"? Is there a technical reason that this is an unsolvable problem?
Considering the increase in computer and network speed over the last decade and a half, it seems strange that this would still be the case. Perhaps it's just that without pressure from competitors there is no pressure on the sites to solve it?
- benjiweber 14y ago1024bit certificates are only valid until December this year. Then we have to use 2048bit, destroying the benefit we get from faster computers. You need SSL accelerators or pretty fast hardware to handle more than a few thousand SSL handshakes per second on a single machine. This is one of the places cost comes from.
- nl 14y agoYou need SSL accelerators or pretty fast hardware to handle more than a few thousand SSL handshakes per second on a single machine. That's not true - or rather it conflates two unrelated problems. 1) HTTPS/SSL (even with larger certificates) isn't computationally expensive: In January this year (2010), Gmail switched to using HTTPS for everything by default. Previously it had been introduced as an option, but now all of our users use HTTPS to secure their email between their browsers and Google, all the time. In order to do this we had to deploy no additional machines and no special hardware. On our production frontend machines, SSL/TLS accounts for less than 1% of the CPU load, less than 10KB of memory per connection and less than 2% of network overhead. Many people believe that SSL takes a lot of CPU time and we hope the above numbers (public for the first time) will help to dispel that.[1] 2) The HTTPS initial handshake is slow. However, this isn't because of the power of the machine, or the certificates size. It's cause by round-trip overheard (See [2]). The solution for that is to reuse HTTPS connections as much as possible (which both your server and the browser should try to do for you anyway. But doing things like using the same hostname can help) [1] http://www.imperialviolet.org/2010/06/25/overclocking-ssl.html http://www.imperialviolet.org/2010/06/25/overclocking-ssl.ht... [2] http://pic.dhe.ibm.com/infocenter/tivihelp/v2r1/index.jsp?topic=/com.ibm.itame2.doc_5.1/ss7aumst18.htm http://pic.dhe.ibm.com/infocenter/tivihelp/v2r1/index.jsp?to...
- benjiweber 14y agoRe-using HTTPs connections is only helpful when you have repeat requests from the same user. It's actually counter productive for tracking beacons or ad impressions, where you rarely see the same user twice in rapid succession. For web applications the overhead is much less significant. Both because you can re-use connections, and because few web applications can manage 20k+ requests per second on low end hardware.
- StavrosK 14y ago> destroying the benefit we get from faster computers. That's precisely the point, though.
- Retric 14y ago2048bit is not all that much harder to encrypt than 1024bit you might go from 1% CPU to 2% CPU but that's about it.
- dangrossman 14y agoI have no background in crypto stuff, but the main issues seem to be (a) SSL virtually requires buying certificates so small/hobby sites won't use it, (b) it involves multiple round-trips and latency is never going to go away unless someone invents faster-than-light communication, and (c) part of what makes encryption work is that it's computationally hard on today's hardware; when it's not, we move to a different algorithm that is, so the computational cost to using encryption versus not using encryption will always be there.
- kbenson 14y ago> part of what makes encryption work is that it's computationally hard on today's hardware; when it's not, we move to a different algorithm that is, so the computational cost to using encryption versus not using encryption will always be there. It's computationally hard to crack. To be useful, it necessarily has to be easier to use than to circumvent.
- reillyse 14y agohttps is no longer "rigid and costly" in a technical sense On a recent website I did my own tests and saw a 2.5% increase in request speed (between http & https). Basically if you aren't a super huge site you aren't going to see any difference. I use https on all my sites, it's just simpler in the long term and you circumvent a lot of problems with security.
- jmillikin 14y ago> SSL virtually requires buying certificates so > small/hobby sites won't use it startssl.com offers free SSL certificates valid in almost every browser, good for one year. I've been using them for my own site with no problems. Certificate cost is no excuse for continuing to use unsecured HTTP. > it involves multiple round-trips and latency is never > going to go away Assuming reasonable protocol design (somewhat problematic for HTTP/1.0, better in /1.1 and SPDY), additional round trips are only a factor for the initial connection setup. Later requests can re-use the existing SSL session. > part of what makes encryption work is that it's > computationally hard on today's hardware Encryption is based on being computationally expensive for a third party. The computational load on the communicating parties is negligible, particularly with modern CPUs. From http://www.imperialviolet.org/2010/06/25/overclocking-ssl.html http://www.imperialviolet.org/2010/06/25/overclocking-ssl.ht... , Google experiences CPU overheads of less than one percent: > On our production frontend machines, SSL/TLS accounts > for less than 1% of the CPU load, less than 10KB of > memory per connection and less than 2% of network overhead.
- stephen_g 14y agoI don't know where the authors got "rigid and costly", but I think it's just BS... 2008 was a long time ago though, so maybe it was slightly more challenging then. When Google went over to SSL for Gmail, they said "On our production frontend machines, SSL/TLS accounts for less than 1% of the CPU load, less than 10KB of memory per connection and less than 2% of network overhead. Many people believe that SSL takes a lot of CPU time and we hope the above numbers (public for the first time) will help to dispel that. " http://techie-buzz.com/tech-news/google-switch-ssl-cost.html http://techie-buzz.com/tech-news/google-switch-ssl-cost.html That, coupled with the general availability of cheap and free CA signed certificates makes the claim pretty baseless.
- nwh 14y agoIs there anybody other than StartCom offering free signed certificated?
- luser001 14y agoSerious question: why do you want it to be free? I paid $10 for an SSL cert. Isn't that low enough?
- dangrossman 14y agoThere are several billion people for which that's significant, whether because of where they live, or their age, or access to credit cards to make purchases online easily. $10 more than doubles the minimum cost of having a website on your own domain.
- jes5199 14y agoan annoying thing about HTTPS is that it requires you to serve each domain name from a separate IP address, and that can be somewhat costly.
- theatrus2 14y agoIf you ignore IE+Windows XP, you can safely use SNI.
- micro-ram 14y agoI would love to...Who's with me? Anyone? Anyone? Bueller?
- cdjk 14y agoAlas, while I feel comfortable not XB testing WinXP because it's less than 1% of users (or whatever threshold you want), I'm not quite as comfortable using that argument with SSL certs. Unfortunately the two solutions (SNI and IPv6) are unlikely to work on XP, so the only hope is to wait until all those computers are replaced or upgraded.
- kelnos 14y ago... and the stock Android browser, IIRC.
- dkulchenko 14y agoFixed since 3.x: https://code.google.com/p/android/issues/detail?id=12908#c13 https://code.google.com/p/android/issues/detail?id=12908#c13
- fpgeek 14y agoTrue, but just like XP, there are lots of pre-3.x devices out there...
- onemorepassword 14y agoThis thread illustrates exactly the problem with HTTPS. The proponents of HTTPS everywhere try to sweep the significant downsides under the rug, whilst others are spreading all kinds of unfounded FUD about HTTPS. The bottom line is that "your mileage may vary". For some applications, SSL is trivial, and there is no excuse not to do it. For other scenarios it's a nightmare with all kinds of undocumented complications. I'm currently working on providing SSL for a SaaS service with various client domains on AWS (i.e., with a limited number of IP-addresses). Doable, but far from trivial or inexpensive.
- robmcm 14y agoIt may not be considered costly when your advertisers pull out because their content is getting replaced via HTTP.
- kevinpet 14y agoHTTPS is not rigid and costly for a single connection. It's rigid and costly because it prevents simple options for caching like sharing google hosted jquery. You can still use a CDN, but then you have more complexity in managing your certificates and pay 1/3 more (Cloudfront pricing). HTTPS is manageable where needed for security. Sleazy ISPs are making it necessary even where security is not a concern -- for example, viewing Apple's homepage.
- cmelbye 14y agoThat's not true. Google seamlessly supports both HTTP and HTTPS, and many other sites do too through simply using // as the protocol.
- iy56 14y agoThey are not making it necessary, only convenient. You're not harmed by seeing a few ads.
- brandon272 14y agoThis point of view assumes that the injected code is harmless and never causes compatibility issues when viewing a website. The moment a JavaScript conflict occurs and Apple loses a sale or the ability to buy a product online is hindered, I think that there will be parties who certainly feel they are being "harmed" by this process of manipulating the content being delivered to end users.
- iy56 14y agoYes, this technique could potentially be misused. You can say the same thing about any technology.
- driverdan 14y agoUsing 3rd party CDNs, like Google Hosted Libraries, are often slower anyway. http://www.stevesouders.com/blog/2013/03/18/http-archive-jquery/ http://www.stevesouders.com/blog/2013/03/18/http-archive-jqu...