HTTPS Everywhere Gets Firesheep Protection

coondoggie writes "The Electronic Frontier Foundation today said it rolled out a version of HTTPS Everywhere that offers protection against 'Firesheep' and other tools that seek to exploit webpage security flaws. Hitting the streets in October, Firesheep caused a storm of controversy over its tactics, ethics and Web security in general. Firesheep sniffs unencrypted cookies sent across open WiFi networks for unsuspecting visitors to Web sites such as Facebook and Twitter, and lets the user take on those visitors' log-in credentials."

And the ISP will sniff you.

[Anonymous Coward]

There's no substitute for end-to-end encryption.

Re:

[tepples]

here's no substitute for end-to-end encryption.

I agree. But practically, even with a source of cheap or free TLS certificates, how does one establish end-to-end encryption for passwords on blogs, forums, and the like without paying extra for a hosting plan that includes a dedicated IP address? Name-based virtual hosting on HTTPS doesn't work with pre-SNI clients such as Windows XP.

Re:

[Lennie]

Don't use passwords, use OpenID ? I guess you can still hijack sessionids, so that is useless.

But SSL/TLS isn't perfect either. Any root-CA or sub-CA can issue any certificate.

We would atleast still need something like DNSSEC to validate what is stored in DNS. So that we can store in DNS, not just the A- or AAAA-record, but also which CA is allowed to sign your certificates.

Even then, if you choose an external CA, it has pretty much been proven that governments can still get certificates from them.

Do you tr

RFC 4398

[tepples]

We would atleast still need something like DNSSEC to validate what is stored in DNS. So that we can store in DNS, not just the A- or AAAA-record, but also which CA is allowed to sign your certificates.

But by the time you're using DNSSEC, the domain registry is already acting as an ersatz CA by signing the CERT record (RFC 4398 [ietf.org]) that you have added to your domain. So I agree that DNSSEC is the real answer to TLS PKI.

Re:

[Lennie]

So far I've not seen anyone implement any specification. This is the only real effort so far:

1. http://www.imperialviolet.org/2010/08/16/dnssectls.html [imperialviolet.org]

2. Dan Kaminsky released Phreebird which includes Phreeload which is a library on top of OpenSSL to verify certificate fingerprints using DNSSEC and a TXT-record.

Re:

[Anonymous Coward]

This is the kinda post someone who does not live outside their own box and doesnt understand how the rest of the world works, and who doesnt understand the the majority of people have zero idea how technology works.

Re:

[Anonymous Coward]

I don't give a rats ass if somebody else in the cafe also wants to know the weather, or also wants to read about Linux concepts...

Don't use unsecured wireless for sensitive stuff.

All stuff is sensitive. Would you like to have e.g. your windows updates guid sniffed and used by some middle east or wherever guys later? Then you=them in terms of tracking by certain agencies, etc.

windowsupdate.microsoft.com: To provide you with the best possible service, Windows Update also tracks and records how many unique machines visit its site and whether the download and installation of specific updates succeeded or failed. In order to do this, the Windows operating system generates a Globally Uniq

Re:

[MachDelta]

B-b-b-but how am I supposed to get on teh intertubes at school? :(

Re:

[Anonymous Coward]

It's actually pretty common, and possibly even the norm.

You can't just use a pre-shared key, so you have to use WPA enterprise. (a PSK is only slightly better than open, for privacy, if everyone knows it, and not terribly useful for regulating access to the network if you only want school affiliates to use the wireless resources).

Often times you can't use the more common EAP types because the authentication data isn't stored in a way that's friendly to your radius servers.

So now you have to write all sorts

Re:

[Anonymous Coward]

Enterprise or Pre-shared key WPA? Pre-shared keys are only marginally better than open, if everyone knows the key. If I know the PSK, I can force you to rekey your session then your traffic is unencrypted to me and I can use firesheep on you.

And the fact that they use "mac-filter" leads me to think it is just PSK.

That isn't to say these mechanisms are completely worthless, but they're not super-valuable.

And I stand by my initial statement -- enterprise WPA in a university setting where you don't manage th

Re:Do Not Use Unsecured Wireless

[bunratty]

It's not as simple as that. The traffic is encrypted only during one part of the way from your computer to the server, so cookies can be sniffed anywhere from the wireless router to the server. But it is as simple as using HTTPS. Then all traffic is encrypted all the way from your computer to the server, and you also have the stronger guarantee that your computer is talking to the server you think it is, so you cookies cannot be sniffed by third parties. StartSSL offers free SSL certificates to allow any site to encrypt all of its traffic.

Windows XP does not support SNI

[tepples]

StartSSL offers free SSL certificates to allow any site to encrypt all of its traffic.

But you will need a separate IPv4 address for each certificate, which usually means a separate IPv4 address for each domain. Will all Windows XP clients be upgraded to an OS that use Server Name Indication [wikipedia.org] before ARIN runs out of IPv4 addresses? I don't think that's likely.

Re:

[yuhong]

To be more precise, XP's built-in crypto library do not support SNI. While IE uses it, I don't think Mozilla uses it, instead they use NSS instead.

Re:

[tepples]

Chrome and Safari also use Windows's built-in crypto library. So unless you want to make your HTTPS site exclusive to Firefox and Opera, or until Microsoft finally shuts down the activation server for XP in 2014, each certificate still needs a separate IPv4 address.

Re:

[Lennie]

Chrome does support SNI on Windows XP, they still use the same Windows certificate store, but I don't know if they hacked around the Windows library or just use a different library.

Re:

[yuhong]

Used a different library:
https://code.google.com/p/chromium/issues/detail?id=43142 [google.com]

Re:

[Lennie]

Thanks for letting me know

Does browsing wikipedia work now?

[CryptoKiller]

Does wikipedia work with HTTPS Everywhere now? I had to disable it because of all the 404 error messages I was getting.

Re:

[somegeekynick]

I did not have problems with accessing Wikipedia, ever. However, the suggestions feature in the search box for Wikipedia as well as Google stopped working ever since 0.2.2 (for me, anyway). 0.9.0 has not rectified this situation.

Duh?

[girlintraining]

Wait, unencrypted signals sent over the air with your password and login is bad? If only someone had told me... /snark

Seriously though: Unencrypted. Open. Network. Come'on guys.

Re:Duh?

[The MAZZTer]

Firesheep never used login credentials. It never needed to. Session cookies were enough to impersonate another user... so any visit to any HTTP page on any site allowed a Firesheep user to impersonate you on that site in theory (of course if you're logged out this is of limited use, but if you're logged in they can impersonate you without login details).

Re:

[leptechie]

The extension forces requests to be sent over SSL/TLS for all communication, as long as the site supports it. Works on Facebook, even Google searches, so yes this is a useful countermeasure. Of course, it is wholly dependent on the site supporting HTTPS in the first place.

I've tried similar extensions, and Facebook gladly connects over HTTPS when manually instructed to, but reverts to normal HTTP on pretty much any click, this just keeps the connection on HTTPS regardless of the link target. The only downsi

Re:

[damaki]

There is a chrome extension working around that. [google.com]

Re:

[moonbender]

Does that work for Google, too? In other words, can you simply use the session cookie sent when performing a Google search to log into a Gmail session? The former is typically http, the latter https. I'm aware that you can use https for searching, too, I'm just wondering. Isn't there some kind of policy that segregates https cookies from http cookies?

Re:

[Bucky24]

Having used firesheep myself, I can tell you that it doesn't work to login to gmail. I was able to sniff my own google account (using a separate computer), but was not able to get past my iGoogle homepage. Attempting to go to gmail prompted a login, even using the firesheeped session cookie. I think I may have tried google docs too, and it also required a login, but I am not sure about that one.

Re:

[Lennie]

Their is also a bug in Chrome when you start up it will connect to the google.com-domain to check for updates and do so over HTTP and the problem with this is, it will also send the normal cookies associated with google.com. Which might give the attacker enough information to get into your igoogle/gmail account.

Re:Duh?

[blueg3]

Many of the sites that Firesheep attacks use HTTPS for their login, so you don't send your credentials in the clear, but fall back to HTTP for delivery of content. The point Firesheep attempts to make is that this is not sufficient -- your unencrypted HTTP requests contain the session cookie that your encrypted login obtained. The session cookie is just as useful, as long as you make use of it "soon".

Probably breaks lots of web sites

[defaria]

Stated simply, many web sites just can't handle https.

Re:Probably breaks lots of web sites

[oodaloop]

Um, no. That would be pretty dumb. IF the site has an https page, it directs to that. If not, it doesn't.

Re:

[budgenator]

HTTPS take more processing power to encrypt and decrypt the traffic, frequently enough traffic flowing through marginally usable server will completely crash and burn if all of the traffic were encrypted; it's like normal traffic would cause a site to be /.ed.

Static vs. dynamic

[tepples]

HTTPS take more processing power to encrypt and decrypt the traffic

This might be a valid concern for static web pages. But the sorts of web sites with which one would use TLS are more dynamic, to the point where they might be called web applications. How much processing power does HTTPS use compared to what the PHP/Python/Perl/Java app and the database use?

Re:

[budgenator]

it's always in addition to what the PHP/Perl/Python/Java uses. On say slashdot you have multiple servers sliced horizontally so each section can be on a separate server if desired as well as vertically so you have web-servers on the outer layer, Database caches in the middle, and the database itself on the inside; or on other sites you have the web-server and database server all on a host shared by ten or twelve virtual hosts. The load that a slashdot can service probably wouldn't be much of a problem going

In addition by how much?

[tepples]

it's always in addition to what the PHP/Perl/Python/Java uses.

But how much addition? Would HTTPS increase the CPU load of a typical PHP blog, forum, or wiki engine by 1%, 10%, 100%, or more?

Re:

[budgenator]

SSL uses strong cryptographic encryption, which necessitates a lot of number crunching. When you request a webpage via HTTPS, everything (even the images) is encrypted before it is transferred. So increased HTTPS traffic leads to load increases. Why does my webserver have a higher load, now that it serves SSL encrypted traffic? [apache.org]

All servers will display an increased load how ever

In my experience, servers that are heavy on dynamic content tend to be impacted less by HTTPS because the time spent encrypting (SS

Re:

[Lennie]

Yes, a higher load, but the load will only be increased by a very, very small marinal number. Just have a look at what the Google study had to say about it.

With the right extension we could even speed up loading of the webpages:

http://www.chromium.org/spdy/spdy-whitepaper [chromium.org]

Because HTTP does not currently do multiplexing of multiple streams over the same TCP (or TCP/SSL) connection. The only solution that is has is to open several connections and because we need to use TCP-slowstart it can't utilize the availa

Re:

[spidr_mnky]

That sounds like a big opportunity for a man in the middle to tell you your site doesn't support ssl. That's still fine against passive listeners, but wouldn't it make more sense to it like flashblock and noscript do? Force ssl by default, and if that doesn't work ask the user if he wants to go with plain http. That would at least give you the opportunity to say, "Gee, gmail usually has ssl ..."

How does HTTPS Everywhere do it?

[Bromskloss]

Does it parse the webpage you are on and rewrite every link to use HTTPS or, better, does it intercept every request Firefox makes and rewrite that before it is sent?

The reason I'm interested is that I want to create an extension that does rewrites in the latter way described, but don't know how to do it.

Re:

[Logic Worshiper]

It's pretty simple, you can often force https simply by typing it in the address bar, if the site has some kind https cert set up, it'll bring you to a secured version of the site. https://twitter.com/ [twitter.com] works and brings to twitter securely, as it does for many mail sites. All you have to do is add a character the html string, which isn't that complicated. For it to really be secure, the server administrator has to secure their site.

Re:

[Kozz]

If you don't find your answer in forums, there's nothing stopping you from looking into the extension yourself, in most cases. Take the XPI file,change the extension to .zip (presuming you're in windows) and extract its contents. View source of the .JS files, and there you have it.

Re:

[Bromskloss]

The latter. They complained that Chrome does not let them hook into the networking component, so a similar add-on is impossible for Chrome. It may rewrite links, too, but that would not protect against external/manually typed in URLs or requests made via Javascript.

Excellent! What is the documentation for learning how to do this? Or, as a backup, where in the code for HTTPS Everywhere is the relevant piece of code?

Re:

[Anonymous Coward]

It does the latter. Requests are intercepted and converted according to pre-defined and user-definable rulesets before being sent.

Re:

[pyster]

If you are using an open wireless you have the same http/https issues everyone else has, regardless of the device you are using.

Re:

[Don_dumb]

Thanks but I am unclear do the apps use http or https to communicate?
Is there any way of knowing what security the apps are using to communicate with the service.
This is important to consider as I haven't seen an iPhone app have an option of securing their connection with remote services. Most people use apps for things like facebook and are entirely at the liberty of the apps' security. There is no 'use https' choice if it doesn't do so.

CA's are the problem, not the crypto

[muckracer]

SSL = Great
SSL + some 600 MITM-Orgs your browser "trusts" = Bullshit

Use HTTPS Everywhere anyway. Great plugin. But forget your much-touted "sense-of-security" because it can't exist in light of the above.

Re:

[Logic Worshiper]

A self signed certificate would be fine for most of what HTTPS Everywhere does.

End to end encryption is for stuff that really matters, not facebook and other crap that's public to the internet anyway.

Re:

[Steeltoe]

You can be fired over someone cracking or spoofing your Facebook identity. I would consider that important.

Who decides what security is important, really?

Re:

[tepples]

A self signed certificate would be fine for most of what HTTPS Everywhere does.

But then how would your users know that your self-signed cert is authentic before installing it into their browsers for the first time? MITM in the wild is a reality [mozilla.org].

Re:

[Lennie]

DNSSEC is the solution for that, but it will take years before we'll have validating resolves in the browser.

In studies it has been shows, even many DSL-routers block DNS over TCP or large DNS-packets.

Re:

[Onymous Coward]

Here's a way of handling certs which doesn't rely on those organizations: Perspectives

http://www.cs.cmu.edu/~perspectives/ [cmu.edu]

Re:

[Lennie]

I don't know and who validates the response from the Perspectives notaries (the agents in the networks which store this information) ?

And what about my privacy ? Do they store the information about the sites I visit (they probably say they don't).

Maybe Certificate Patrol is a good start:

https://addons.mozilla.org/en-US/firefox/addon/6415/ [mozilla.org]

It works like SSH, everytime you connect it will tell you if something changed since you last connected. It also will tell you what changed.

Re:

[muckracer]

I can't wait for CAsheep....

Actions you must take for firesheep protection

[Fnord666]

According to the release notes, there are specific actions that you must take to enable some of this protection:

The 0.9.0 release of HTTPS Everywhere is a new beta version designed to offer improved protection against Firesheep. Most notably, it can provide much better protection for Facebook, Twitter and Hotmail accounts, as well as completely new protection for bit.ly, Dropbox, Amazon AWS, Evernote, Cisco and Github. Unfortunately, in order to obtain maximum Firesheep protection, especially on Facebook, you must take two extra steps:

• Turn on the "Facebook+" rule. You can do that in the Tools->Add Ons->HTTPS Everywhere->Preferences menu. It isn't on by default, because it can cause Facebook Apps to raise errors. We're still waiting for Facebook to fix this, and the chat problem :(.
• Install the Adblock Plus Firefox extension too, and use it to block the insecure http:/// [http] adds and trackers that Facebook (and other sites) sometimes include.

Re:

[pyster]

The release notes maybe incorrect, or dated, as it was on by default for me. the amazon rule is not, and has (buggy) next to it.

Secure cookies

[fulldecent]

Assume sites want to prevent firesheep, and do not want https everywhere. Does secure cookies fix this?

Login via HTTPS, get secure cookie ("the token") . Then on each page load, use this token to sign your request.

This can be done with existing technology, but requires Javascript.

Re:

[Mark Hood]

It can be done, but it's not being done - that's why this happens.

Why not HSTS?

[Anonymous Coward]

The HTTP Strict-Transport-Security (HSTS) header and its predecessors, X-Force-HTTPS and X-Force-TLS, enable HTTP sites to declare that and how they want to be accessed over a secure connection.
The HSTS header is not recognized by Firefox 3.x. Firefox 4 supports it but without an UI. The extensions ForceTLS and STS UI deal with that, respectively.
These extensions should be merged with HTTPS Everywhere. It's unreasonable to expect people to manually enter all the sites they use, and it's equally unreasonable

ipod

[religious freak]

I need to look this up, but does anyone know how to use this on an unjailbroken ipod, or how about the facebook application on the ipod?

I know the dangers and concerns, but I still use unencrypted wifi like all those that don't even have a clue. I suppose I'm the worst of all... but I bet I'm not alone. It really is amazing how a system with so many vulnerabilities manages to stay together and grow for decades :-P

https://twitter.com almost works...

[yuhong]

https://twitter.com/ [twitter.com] almost works, but I sniffed the packets using Wireshark and unfortunately they still make one HTTP request, which because the session cookie is not marked secure is sent insecurely along with it. I remember reading that it was made using XMLHTTPRequest.

Re:

[yuhong]

Ah, I think that was from days ago. I just checked today using Wireshark again and they fixed it, which means https://twitter.com/ [twitter.com] should be safe now.