Security is a people problem. OK, you already know that. But last week, the SANS Institute finally recognised it too, in its list of the top 20 Internet security risks of 2007. Topping the chart of new, hard-to-defend-against risks were vulnerabilities in custom web applications and "gullible, busy, accommodating computer users, including executives, IT staff and others with privileged access".
According to the SANS report, cybercrooks are still running their automated attack programs, looking for security holes in unpatched and misconfigured software (average time until an attack after a new system is attached to the internet: five minutes).
But IT shops are now much better at blocking those attacks with technology. So attackers are getting more specific. They're launching tightly targeted phishing attacks against specific users as a way to get at corporate and personal information, including customer credit card data and individual online banking passwords.
In other words, they're not just going after vulnerable software. They're aggressively aiming at vulnerable people, too.
It's great news that SANS is publicly recognising security as a people problem.
Not such great news is what SANS recommends we do about it. Option 1: Launch our own test attacks against users, and cut off internet access for those who fail the test. Option 2: Dramatically beef up our monitoring and forensics systems, so we can constantly search for intrusions.
Those are classic security responses. The first one is wrongheaded, and both of them are naive.
What's wrong with staging our own phishing attacks? Nothing, really. But cutting off users who fail the test? That turns us into the enemy of our users. Instead of encouraging them to be vigilant against outside attackers, it will guarantee that IT gets zero trust, zero cooperation and zero help from users.
And that animosity won't stop with the security group. The IT people who take the brunt of it will be those on the helpdesk and development teams and anyone else who deals directly with users.
Worst of all, it won't work. How long will a top sales guy remain without internet access after he fails the test? No time at all. When it's between business and security, the choice will go to business every time. There will be no consequences. And IT's great new security programme will be a fiasco.
What about beefing up our monitoring and forensics? Sure, that's a great idea — it won't antagonise users and is certain to improve our chances against attackers. Just one question: How will we pay for expensive new security systems that generate no revenue and might not be needed?
Answer: We won't. We'll never get budget approval for a proposal so focused on security and so naive about business reality.
Understanding that security is a people problem is important. But it's not enough — just as seeing security as a technology problem wasn't enough.
We have to take the next step and make the case for security as a business problem.
As long as security is seen as just a cost — or, worse, as the enemy of business — we'll never have CEO support for critical security initiatives.
We have to frame security as a business enabler, not a business crippler. We have to sell real benefits: safer customers, more reliable systems, executives whose bank accounts aren't emptied.
Then we can get the budgets for beefed-up monitoring. And the cooperation of business-side managers to keep employees in line.
And most of all, confidence from top-level management that we really are looking out for the business.
Security is a business problem. You know that. Now let's go to work on getting the rest of the business to recognise it too.