Signed, sealed, delivered: by who?
There is a safe and easy way to guard against phishing emails, yet 58% of us receive one every day. So why we don't make the most of digital signatures?
You've got an email from your bank. Or at least, it says it's from your bank. But what if it's a fake? Chances are it is: up to 150m phishing emails are sent every day, according to the Anti-Phishing Working Group (APWG). Sophos estimates that 58% of email users receive at least one each day and 22% get five or more. And people fall for them: figures from Apacs, the clearing banks' organisation, found that in the first half of 2008, online banking fraud in the UK increased by 185% to £21.4m.
But if the email is really from eBay or PayPal, it will have been digitally "signed" - that is, end with a string of digits generated from a private cryptographic key and the original message, using DomainKeys Identified Mail (DKIM), an emerging standard created by Yahoo and Cisco in 2008 after years of deliberation. An agreement with Yahoo and Gmail means users of those email services won't receive messages purporting to be from either organisation if they lack DKIM signatures. They won't even make it as far as the spam folder.
Digital signatures sound like a formidable weapon against phishing, since in theory they give users a way to check that emails are genuine, and they're less likely to open the fake ones. Ever since Pretty Good Privacy (PGP) arrived in 1993, people have had a way to cryptographically sign their own emails, as have organisations - you can create your own cryptographic key for free. We could all be free of phishing; yet we aren't. The question is, why not?
In 1999 Jim Bidzos (currently the interim chairman and president of VeriSign) told the RSA data security conference: "Public-key cryptography is a solution in search of a problem, and e-commerce is that problem." In 2000, the Guardian reported that public-key infrastructures (PKIs) could be used to ensure electronic communications are trustworthy. But in 2009, PKI is still waiting for its heyday.
"PKI is everywhere and most of us don't realise," says John Bullard, global ambassador for the identity authentication system IdenTrust. "But it's predominantly used in the corporate world where transaction values are higher. For example, your direct debits and salary are wrapped up in a PKI-based digital signature when they're sent to Bacs (the banks' automated clearing service)."
Keys to the problem
But it hasn't reached the user level yet. One problem is the number of different encryption systems available. There's DKIM, as well as the commercial PGP, its open-source alternative Gnu Privacy Guard (GPG), S/MIME and more. "We're in an era where there are a lot of PKI systems," says Bullard. "Individual solutions create little islands of trust. The problem comes when you try to go from one island to the next."
Garreth Griffith, PayPal UK's head of risk management and a director of Get Safe Online, says the competition between digital signature technologies has delayed their implementation. "The battle between companies to figure out which one to adopt created a time-lag," he says. "Should you build for all of them or wait for one to become dominant? When Cisco and Yahoo came together to create DKIM, that was a huge step forward as people could start to back someone and develop accordingly."
So why aren't digital signatures being used more widely? "Mail senders, providers and ISPs need to implement their parts of the systems, such as Google's deal with eBay and PayPal," says Mark Bowerman, a spokesman for Apacs. "PKI is very secure, but it's complex and slow to roll out to large numbers of people. The take-up by mail providers to support PKI has been relatively low to date, so the benefits are not yet as apparent as they could be."
Some email clients that support signed mail add an icon, such as a ribbon, to signed messages, or offer an option to view more details, but many providers and clients don't display digital signatures. "I haven't seen any of the major email clients do much with the information, even if it's there in the first place," says Adam Geller, vice-president of enterprise and government authentication for VeriSign.
"The technology is very capable, but the way it's being exposed to end-users leaves something to be desired. If I receive an email from PayPal that is not officially signed in a way I can validate, I should assume that email is suspicious. But your email client may not show you that in a meaningful enough way."
At the Anti-Spam Research Group (ASRG), chairman John Levine agrees: "Email providers should be doing it. The better we can tell email is real, the easier it is to build a case against the guys selling the fake stuff. The vast majority of big companies buy email from specialist vendors who say yes, this is really important and they'll do it in the next version. The big providers know about it and the technology is quite sound, but the software vendors are slow."
Lack of support
According to Jon Callas, chief technology officer for PGP Corporation and one of the authors of DKIM, email providers are reluctant to let end-users see digital signatures because they don't want to deal with the ensuing support calls asking what they are.
"The large ISPs wanted to be able to make DKIM completely invisible because they didn't want to get tech support calls. Yahoo wanted to show their users, others don't. If 1% of their customer bases called in, it would cost enough that they wouldn't want to do it," he says.
And that's the crux: the cost of phishing emails falls on the person who receives it, not the organisations that let it through, whereas the cost of those support calls falls firmly on the ISPs; so even if implementing signing would lead to an overall fall in fraud, for which ISPs' customers (and banks) could be grateful, there's nothing in it immediately for the ISPs.
For the mail that doesn't get filtered, end-users need to be able to check whether or not it's signed. After all, there's no point adding digital signatures to emails if nobody bothers to look for them. "Criminals play on trusted brands," says PayPal's Griffith. "You see a word you trust and you drop your defences. It takes time to educate end users to look for that seal and know what it means, just as it's taken years to educate consumers about looking for a lock on a secure website. The boom of web-based email programs makes it even more complicated."
Even so, Callas says progress is being made. "Email providers are starting to support digital signatures, but it's a gradual process. It's starting to work - early adopters are seeing a drop-off in phishing attempts."
However John Colley, managing director (Europe, Middle East and Africa) of (ISC)2, which educates and certifies security professionals, and formerly head of information security at the Royal Bank of Scotland, says digital signatures aren't a viable solution. "The use of digital signatures as a means of verifying emails to customers has been widely assessed as impractical," he says. "The registration and key maintenance process to verify encrypted emails would have to be as effective as the process of applying for a passport, putting the costs too high for vendors that cater to the wider public."
Unpicking the lock
Colley says there is also too much required of the non-technical user. "They are not going to make the effort without a clear understanding of the value they get for it, which requires an awareness exercise - better to just educate on the phishing risk itself."
The security expert Marcus Ranum, who invented the proxy firewall, says PKI won't work because there are too many other risks. "Even if there was a global public-key infrastructure, there would immediately be a global black market for stolen keys to be used for spamming," he says. "The keys would probably be a file, encrypted, sitting on a user's hard drive. That file, and the encryption password, would be stealable by a Trojan horse or spyware. We'd be right back where we are now."
So the advice remains the same: don't click on the link in the email from your bank. Go to the website via your browser. PKI may be the solution to a problem that's been identified, but the problem is that nobody can be persuaded to pay for it.
Anne Wollenberg
The Guardian, Thursday 8 January 2009
guardian.co.uk © Guardian News and Media Limited 2009
There is a safe and easy way to guard against phishing emails, yet 58% of us receive one every day. So why we don't make the most of digital signatures?
You've got an email from your bank. Or at least, it says it's from your bank. But what if it's a fake? Chances are it is: up to 150m phishing emails are sent every day, according to the Anti-Phishing Working Group (APWG). Sophos estimates that 58% of email users receive at least one each day and 22% get five or more. And people fall for them: figures from Apacs, the clearing banks' organisation, found that in the first half of 2008, online banking fraud in the UK increased by 185% to £21.4m.
But if the email is really from eBay or PayPal, it will have been digitally "signed" - that is, end with a string of digits generated from a private cryptographic key and the original message, using DomainKeys Identified Mail (DKIM), an emerging standard created by Yahoo and Cisco in 2008 after years of deliberation. An agreement with Yahoo and Gmail means users of those email services won't receive messages purporting to be from either organisation if they lack DKIM signatures. They won't even make it as far as the spam folder.
Digital signatures sound like a formidable weapon against phishing, since in theory they give users a way to check that emails are genuine, and they're less likely to open the fake ones. Ever since Pretty Good Privacy (PGP) arrived in 1993, people have had a way to cryptographically sign their own emails, as have organisations - you can create your own cryptographic key for free. We could all be free of phishing; yet we aren't. The question is, why not?
In 1999 Jim Bidzos (currently the interim chairman and president of VeriSign) told the RSA data security conference: "Public-key cryptography is a solution in search of a problem, and e-commerce is that problem." In 2000, the Guardian reported that public-key infrastructures (PKIs) could be used to ensure electronic communications are trustworthy. But in 2009, PKI is still waiting for its heyday.
"PKI is everywhere and most of us don't realise," says John Bullard, global ambassador for the identity authentication system IdenTrust. "But it's predominantly used in the corporate world where transaction values are higher. For example, your direct debits and salary are wrapped up in a PKI-based digital signature when they're sent to Bacs (the banks' automated clearing service)."
Keys to the problem
But it hasn't reached the user level yet. One problem is the number of different encryption systems available. There's DKIM, as well as the commercial PGP, its open-source alternative Gnu Privacy Guard (GPG), S/MIME and more. "We're in an era where there are a lot of PKI systems," says Bullard. "Individual solutions create little islands of trust. The problem comes when you try to go from one island to the next."
Garreth Griffith, PayPal UK's head of risk management and a director of Get Safe Online, says the competition between digital signature technologies has delayed their implementation. "The battle between companies to figure out which one to adopt created a time-lag," he says. "Should you build for all of them or wait for one to become dominant? When Cisco and Yahoo came together to create DKIM, that was a huge step forward as people could start to back someone and develop accordingly."
So why aren't digital signatures being used more widely? "Mail senders, providers and ISPs need to implement their parts of the systems, such as Google's deal with eBay and PayPal," says Mark Bowerman, a spokesman for Apacs. "PKI is very secure, but it's complex and slow to roll out to large numbers of people. The take-up by mail providers to support PKI has been relatively low to date, so the benefits are not yet as apparent as they could be."
Some email clients that support signed mail add an icon, such as a ribbon, to signed messages, or offer an option to view more details, but many providers and clients don't display digital signatures. "I haven't seen any of the major email clients do much with the information, even if it's there in the first place," says Adam Geller, vice-president of enterprise and government authentication for VeriSign.
"The technology is very capable, but the way it's being exposed to end-users leaves something to be desired. If I receive an email from PayPal that is not officially signed in a way I can validate, I should assume that email is suspicious. But your email client may not show you that in a meaningful enough way."
At the Anti-Spam Research Group (ASRG), chairman John Levine agrees: "Email providers should be doing it. The better we can tell email is real, the easier it is to build a case against the guys selling the fake stuff. The vast majority of big companies buy email from specialist vendors who say yes, this is really important and they'll do it in the next version. The big providers know about it and the technology is quite sound, but the software vendors are slow."
Lack of support
According to Jon Callas, chief technology officer for PGP Corporation and one of the authors of DKIM, email providers are reluctant to let end-users see digital signatures because they don't want to deal with the ensuing support calls asking what they are.
"The large ISPs wanted to be able to make DKIM completely invisible because they didn't want to get tech support calls. Yahoo wanted to show their users, others don't. If 1% of their customer bases called in, it would cost enough that they wouldn't want to do it," he says.
And that's the crux: the cost of phishing emails falls on the person who receives it, not the organisations that let it through, whereas the cost of those support calls falls firmly on the ISPs; so even if implementing signing would lead to an overall fall in fraud, for which ISPs' customers (and banks) could be grateful, there's nothing in it immediately for the ISPs.
For the mail that doesn't get filtered, end-users need to be able to check whether or not it's signed. After all, there's no point adding digital signatures to emails if nobody bothers to look for them. "Criminals play on trusted brands," says PayPal's Griffith. "You see a word you trust and you drop your defences. It takes time to educate end users to look for that seal and know what it means, just as it's taken years to educate consumers about looking for a lock on a secure website. The boom of web-based email programs makes it even more complicated."
Even so, Callas says progress is being made. "Email providers are starting to support digital signatures, but it's a gradual process. It's starting to work - early adopters are seeing a drop-off in phishing attempts."
However John Colley, managing director (Europe, Middle East and Africa) of (ISC)2, which educates and certifies security professionals, and formerly head of information security at the Royal Bank of Scotland, says digital signatures aren't a viable solution. "The use of digital signatures as a means of verifying emails to customers has been widely assessed as impractical," he says. "The registration and key maintenance process to verify encrypted emails would have to be as effective as the process of applying for a passport, putting the costs too high for vendors that cater to the wider public."
Unpicking the lock
Colley says there is also too much required of the non-technical user. "They are not going to make the effort without a clear understanding of the value they get for it, which requires an awareness exercise - better to just educate on the phishing risk itself."
The security expert Marcus Ranum, who invented the proxy firewall, says PKI won't work because there are too many other risks. "Even if there was a global public-key infrastructure, there would immediately be a global black market for stolen keys to be used for spamming," he says. "The keys would probably be a file, encrypted, sitting on a user's hard drive. That file, and the encryption password, would be stealable by a Trojan horse or spyware. We'd be right back where we are now."
So the advice remains the same: don't click on the link in the email from your bank. Go to the website via your browser. PKI may be the solution to a problem that's been identified, but the problem is that nobody can be persuaded to pay for it.
Anne Wollenberg
The Guardian, Thursday 8 January 2009
guardian.co.uk © Guardian News and Media Limited 2009
