IT Admins Must 'Think Like Hackers'

Veteran IBM security architect Jeff Crume on Monday urged IT administrators to start thinking like malicious hackers to fully understand the ways in which a corporate network can be breached.

In a standing-room-only presentation at the InfoSec World conference here, Crume identified a long list of weaknesses targeted by attackers and recommended that businesses look beyond traditional firewalls, filter rules and anti-virus software.

"If you're a security administrator these days, it's not a good time to be sleeping. The attacks these days are more voluminous and the dynamics aren't working in our favor," Crume said, pointing out that attack mechanisms haven't changed much over the last five to 10 years.

"There's nothing more dangerous than presumed security," Crume warned. "There are some things you think you know but they come with many misconceptions."

First off, Crume said, the overdependence on perimeter firewalls gave businesses a false sense of security because that first line of defense can become the single point of failure. "Almost all attacks that are troublesome, firewalls can hardly prevent," he said. "The hacker doesn't want you to know that once he breaks past the firewall, he's in the network and can do whatever he wants."

"A firewall can't detect many types of attacks and it can't tell if a particular packet is malicious. A firewall can't defend against attacks that don't go through it, so if it's only at the perimeter, there's a sitting duck aspect to it," he said.

Even worse, Crume argued, a lot of malicious attacks have historically come from within an organization, making perimeter firewalls largely meaningless.

Interestingly, Crume recommended that business use more firewalls to fix the problem of overdependence. Because the "bad guys" can be sitting within the network, he suggested that businesses carve off separate security zones internally to block or limit access to parts of the network.

"If you look at how some worms propagate, it looks like an insider attack. If someone goes on the road with a laptop and picks up a Trojan, he is a risk to the corporate network when he returns. These are the types of internal risks that the firewall can't block," he said.

"[With] all the big worm attacks, if you look closely, the exploits start from outside but the big spread happens within the network," Crume said, likening the protection mechanism to the shutoff valve used by plumbers to minimize the damage from a burst pipe.

"The outsider is probing and stumbling around like a thief in the night. But the insider knows exactly where the data is and he knows his way around the intrusion detection system. He is a bigger threat to do damage unless you do a better job of implementing the separation," Crume said. "If the intranet holds the keys to the kingdom, you need to treat it as a semitrusted network."

Humans are the weakest link
Another vulnerability hackers prey on, Crume warned, is that humans are the weakest link in the security chain. "We can create all the technological defenses but nothing can stop an employee or user who wants to subvert the process. The human who doesn't understand or know what the process is is a big, big problem."

Crume's presentation included a segment on social engineering, the concept of tricking an authorized person into revealing IDs, passwords and other confidential information.

Social engineers use low-tech hacking methods to collect basic data that can be used to persuade insiders to give up passwords. In some cases, hackers go "dumpster diving," Crume said, explaining that an organization's trash often contains important clues for social-engineering tricksters.

"You can go through a dumpster to get a corporate phone directory which, to a social engineer, is gold. He can get names, titles and responsibilities of everyone in the company from that directory. He can assume the identity of someone on that list and do all kinds of dangerous things," Crume said.

The password problem
Another nightmare for the enterprise security admin is the inherent weakness that comes with depending on passwords. "Passwords aren't secure. We know it's a problem but we're not doing anything about it. As an industry, we haven't gotten better in the last ten years in dealing with this problem," Crume said.

For one thing, employees within the network use trivial passwords that can be easily guessed or cracked by readily available programs. "If your password is about you, it's not a secret. If it's based on public information about you, it's probably likely that others will know. You then become a big target for a social-engineering attack."

He said businesses have tried implementing complicated password-creation schemes but in those cases the employees simply scribbled the passwords on yellow sticky notes and pasted them on the computer monitors.

Crume described using the LophtCrack tool to help a friend who had lost his password. "Some claim a 30-percent success rate with this program, which is scary."

With LophtCrack, a hacker can launch dictionary-based attacks in multiple languages if trivial passwords are used.

Crume recommended businesses invest in newer technologies like single sign-on tokens, smart cards or biometrics authentication. "These aren't the holy grail of security and I don't want to oversell what they can offer. But, right now, they can help solve a lot of identity management-related problems."

Among other things hackers take advantage of, Crume said, are unprotected rogue Wi-Fi networks set up by employees; the availability of easy-to-use worm generators; the increased discovery of down-level software flaws; and the fact that virus defenses are largely inadequate.

Source
 
Back
Top