When it comes to securing the enterprise, the attackers have the advantage. Defenders are required to protect against every conceivable threat while the attacker needs only a single attack vector to penetrate a network.
The universe of potential intrusion vectors is vast: faulty authentication mechanisms, gaps in the perimeter network, legacy applications, and, of course, human behavior are just a few examples. Unfortunately, enterprise security teams tend to focus on a handful of information security domains:
- Patch management and 0-day threats
- Malware and endpoint protection
- Network security
"Network security" has come to be synonymous with "perimeter security." Secure the perimeter, the thinking goes, and everything in the datacenter can operate in an environment of mutual trust. Combined with strong authentication mechanisms, this produces a comfortable, low-maintenance state of affairs. Securing only those systems that face the Internet is a whole lot easier than securing the thousands of servers in the datacenter.
Unfortunately the perimeter is but one attack vector of thousands. As breach after breach hits the news, security professionals have realized that securing the perimeter is not enough. And with that acknowledgment, they are now slowly proceeding through the five stages of security grief.
Denial. In this stage, you, the security pro, believe you can’t be breached. Your DMZ is locked down, your stakeholders comply with your policies, and you’ve bought an intrusion detection system (or three). Your job, then, is pretty easy: keep the firewalls up to date, scan the alerts every morning, quarantine the occasional malware infection, sleep well at night. Other organizations are getting breached. What’s wrong with their security people? They probably don’t have an IDS.
Anger. Slowly, an uncomfortable reality dawns: the average time-to-detection for an enterprise breach is somewhere around four months. New attack vectors emerge as headlines in the news. The perimeter is secure, but your contractors and business partners have access to your network, so you’re only as secure as they are. Your favorite restaurant/ department store /multinational bank gets hacked, and you spend an afternoon updating all your recurring payments with your new credit card number. Meanwhile, the number one malware delivery vector is phishing emails. Still.
Bargaining. In search of answers, you turn to the booming infosec industry. There are so many products. So many! You buy them all. The cost of a breach far outweighs your minimal savings in neglecting to buy the one product that would have prevented that breach. You switch on all the endpoint protection you can find. You log everything; your SIEM bursts with event data. You get thousands of alerts every day. After a while you stop reading them.
Depression. Another breach hits the news. One of your vendors proudly announces their product alerted on the breach, which went undetected for four months as attackers siphoned data out of the fortress. You think about the vendor, with whom you’ve spent ungodly sums. Will they detect your breach? You think about the company that got hacked. Are you better at your job than they are? Is it even possible to be good at this? Is it possible to be good at anything? You resolve to get drunk.
Acceptance. The next morning, your head is pounding. You sit down at your desk and unlock your computer. Suddenly, a thought: it’s no longer about whether you’ll get hacked, and it’s not even about when. You realize there might be attackers roaming your network right now and you wouldn’t know about it for months.
So what comes next?
The worst that can happen, you reason, is you get fired. But when the standard for breach discovery is so low, all you have to do is detect an intrusion faster than the other guy.
You study defense-in-depth. You deploy datacenter-level visibility. You monitor for DNS exfiltration, SSL exfiltration, HTTPS exfiltration. You deploy machine learning for anomaly detection. You audit your partners’ security practices and lock down the partner network. There’s no magic bullet. You ignore the alerts and start hunting for threats.
You haven’t been breached yet, but you find all sorts of problems. Adware is everywhere. Your network segments are too broad, allowing for plenty of lateral movement. Software developers are logging in to production databases using privileged credentials. Your internal firewalls are passing all sorts of traffic. Pretty much anybody can access the storage systems.
Finally, data you can use! One by one, you lock down internal attack vectors. You microsegment your applications and deploy next-generation firewalls within the datacenter. You implement two-factor authentication and continuously monitor compliance. You have three columns of Post-Its on your whiteboard: the security hygiene measures your organization needs, the ones it already has, and the ones you’re monitoring. Gradually, the Post-Its move to the right.
You keep hunting the network for threats. Another breach hits the news: the attackers lurked in the network for three years. You think about the security teams at that company. Are you better at your job than they are?
You accept it: Of course you are.