Encrypted e-mail: How much annoyance will you tolerate to keep the NSA away?
How to to encrypt e-mail, and why most don’t bother.
The uses of asymmetry
The full extent of the cooperation between the NSA and various technology companies is unclear. It will probably remain that way for the foreseeable future. For the time being, however, it seems likely that the standard cryptographic tools used to secure data “in flight”—that is to say, the SSL that protects data traveling between machines on the Internet—remain secure as long as certain best practices are used.
That protects against some threats, such as wholesale monitoring of Internet traffic of the kind the NSA is known to engage in, but it doesn’t do anything to protect data that’s “at rest.” That is to say, SSL doesn’t do anything to prevent a company like Google or Microsoft from handing over an archive of your e-mail in response to a court order. The e-mails are just lying around on some Google server somewhere.
If you don’t want a government, service provider, employer, or unauthorized party to have access to your mail at rest, you need to encrypt the mail itself. But most encryption algorithms are symmetric, meaning that the encryption key serves a dual purpose: it both encrypts and decrypts. As such, people encrypting mail with a symmetric key would be able to decrypt other mail that used the same symmetric key. While this would protect against anyone without the key, it wouldn’t be very useful as an encrypted e-mail system.
The solution to this is asymmetric cryptography. In asymmetric encryption there are two opposite keys, and a message encrypted with one key can only be decrypted with the other. The two keys are known as a private key, which as the name might suggest is kept private, and a public key, which is broadcast to the world. Each time you want to send an e-mail to someone, you encrypt it with the recipient’s public key.
Asymmetric encryption is also used to perform mail signing. For this, the mail sender encrypts a hash, or mathematical fingerprint, of their file, producing a signature. Hashes are designed so that any small change to the message’s text will produce a different hash value. Anyone reading the mail can then decrypt the signature using the sender’s public key, giving them the original hash value. They can then compute the hash value of the mail they received and compare the two. If the values are the same, the message hasn’t been modified. If they’re not, it has—and we’ll see the uses of this later on.
Making things even more complex, having encryption support isn’t itself enough. To a great extent, you don’t control the things that are in your own inbox. That’s all mail that someone else has sent you. If you want your inbox to contain encrypted mail that only you can read, you need to be sure that people sending you mail are encrypting that mail when they send it. And if you want to be sure that everything in your sent mail folder is encrypted, you’ll need to send other people encrypted mail.
As a result, e-mail encryption is not something you can impose unilaterally. To protect the contents of your account, you need to ensure that everyone you communicate with is in a position to handle encrypted mail—and is willing to use that ability.
Finally, e-mail encryption doesn’t encrypt everything. Certain metadata—including e-mail addresses of both sender and recipient, time and date of sending, and the e-mail’s subject line—is unencrypted. Only the body of the mail (and any attachments) gets protected.
If you’re happy with these constraints, e-mail encryption is for you. Unfortunately, it can be complicated to use.
Cutting through the complexity
Few e-mail programs have PGP encryption features enabled by default. And even if they do, end users must still navigate a series of mazes that are long and confusing. Tasks include generating the key pair that will lock and unlock the communications and storing the private key in a location where no one else can get it. It also requires securely sharing a public key with every single person who wants to send you a private e-mail and securely getting a unique public key from each person you want to send encrypted e-mail to. No wonder most people—reportedly including Glenn Greenwald, the Guardian reporter who exposed aspects of the secret NSA dragnet—need time getting up to speed.
Fortunately, free e-mail encryption programs are available for all major operating systems, and the ability to use them effectively isn’t out of the grasp of average computer users if they know where to look. What follows is a set of step-by-step instructions for using GnuPG, the open-source implementation of the PGP encryption suite, to send and receive encrypted e-mails on machines running Microsoft Windows and Mac OS X.
After that, we’ll show readers how to use a similar crypto standard called S/MIME, which may prove simpler to deploy because it is already built into many desktop and mobile e-mail clients, including Outlook and Thunderbird. (Interested in S/MIME? Skip directly to page three.)
Linux will be touched on only briefly because much of the functionality is already included in various distributions and because many Linux users already have PGP down cold. (Users are invited to provide Linux instructions and screenshots in the comments following this article.)
PGP on Windows
The basic element you’ll need to encrypt mail is software to generate and manage your key pair and make them work with whatever e-mail program you happen to use. On Windows, there’s no shortage of proprietary apps that will do both, with Symantec’s PGP Desktop E-mail being perhaps the best known. There’s nothing wrong with this offering, but it’s almost $200 for a single-user license. This tutorial will instead focus on the open-source Gnu Privacy Guard, which is available for free on Windows, Mac, and Linux platforms.
GnuPG, or simply GPG, is still available mostly as a command-line tool, meaning there’s no graphical interface many end users would feel more comfortable using. Rather than learn a long list of GPG commands, many e-mail users are better off installing graphical implementation of GPG. On Windows, Gpg4win will give you everything you need to generate strongly encrypted messages that can be sent and later decrypted by the intended receiver using standard e-mail programs.
At time of writing, the most recent version of Gpg4win is 2.1.1 and it’s available here. After downloading such a sensitive piece of software you’ll want to confirm the installer hasn’t been tampered with and truly came from Gpg4win rather than a site masquerading as gpg4win.org. To do that, we’ll need to check the SHA1 checksum for the downloaded file and make sure it matches the hash—a94b292c8944576e06fe8c697d5bb94e365cae25—listed on the Gpg4win download page. For those who prefer a graphical interface, use HashCalc. Install HashCalc and then open the program. In the “data” box, navigate to the folder where the downloaded gpg4win-2.1.1.exe file is located. In our case, since the SHA1 hash calculated by HashCalc matches the SHA1 digest provided on the Gpg4win download page, we have a high degree of confidence the file we’re about to install is genuine.
For readers who prefer command lines, Microsoft’s File Checksum Integrity Verifier may be a better way to check the SHA1 hashes. You’ll need to download and extract the FCIV package and follow the instructions in the readme text file, including making sure the folder containing the FCIV executable file has been added to the system path of Windows. With that out of the way, open a Windows command window and navigate to the folder containing the Gpg4win installer.
Once you’re sure you have the real gpg4win-2.1.1.exe, double-click on the file and click Yes to the User Access Control dialogue. When presented with the Gpg4win installation welcome screen, click Next, and then click Next at the following window to accept the Gpg4win license agreement. The next screen will allow you to choose the precise GPG components you want to install. Make sure you install all available components, including GPA, which is short for the GNU Privacy Assistant. Click Next at the Choose Components screen and again at the Destination and Install Options screens.
At the Install Options screen, makes sure the “start menu” box is checked, click Next, and at the next window click Install. We won’t be using S/MIME for now, so if you see any screens referring to Trustable Root Certificates, you can click the box to skip configuration and click Next. The installation is now complete.
When you click on your Start menu and choose All Programs, you should now see a Gpg4win folder. Highlight it and choose GPA. This is the GNU Privacy Assistant. We’ll use it to generate our key pair, and later we’ll use it to store the public keys of people who will receive our encrypted messages. The first time you open GPA, you’ll see a screen asking if you want to generate a private key. That’s exactly what we want to do, so click “Generate key now.”
In the screens that follow, enter your name and e-mail address. When asked if you want to back up your key, choose “Do it later.” It’s not that this step isn’t important, but we’ll want to back up the key only after we’re satisfied that we’ve done everything correctly. Next, you’ll need to choose a passphrase to protect your key. Your passphrase is like the password protecting an e-mail or Web account. Except rather than preventing an unauthorized person from accessing your account, it prevents the person from using your private key should it ever be lost or stolen. In other words, the password is extremely sensitive. It should have a minimum of nine characters, but 18, 27, or even 36 characters are even better. For more tips on generating a strong password, see Ars Senior Reporter Jon Brodkin’s discussion of master passwords here. When you’re finished, you’ll have generated your first key pair: the public key you will share with other people so they can send encrypted messages that only you can read, and the private key you’ll use to decrypt those messages.
While generating your key, be sure to set an expiration date, rather than allowing it to remain valid forever. This way, keys that new users abandon, lose or never end up using won’t remain on public servers indefinitely. Remember also to backup your private key somewhere that’s extremely safe. Storing it on a USB stick that’s stored in lock box is one suitable method. You may also want to upload your public key to one or more public key servers. These servers give crypto users a way to make their keys available to others and to fetch other people’s public keys.
Now that we’ve generated our first key pair, let’s import the public key of someone else so we’ll have it later when we’re ready to send them our first encrypted e-mail. For this, get someone to give you their public key, preferably in person. It will look something like this:
-----BEGIN PGP PUBLIC KEY BLOCK-----
Version: GnuPG v2.0.17 (MingW32)
-----END PGP PUBLIC KEY BLOCK-----
Take the public key of a real-world contact and save it to a file named something like key.txt. If you don’t have a real-world contact who has a public key, save the above public key to a file and name it key.txt. Now, with GPA open, choose the “Import” icon, navigate to the disk location of key.txt, highlight the file, and click Open. Congratulations. You’ve just imported your first public key. Don’t get too excited just yet. You’ll need to import a public key for each person you want to send encrypted mail to.