Problem solve Get help with specific problems with your technologies, process and projects.

How to defend against pivot attacks in the enterprise

Nick Lewis explains what a pivot attack is and tells how to go about defending your organization against this threat.

Industry expert Jeremiah Grossman recently wrote a piece on his blog about pivot attacks. What are pivot attacks,...

and how can we prevent them from affecting our organization?

In a pivot attack, an attacker targets a lower security host, where less security is in place, which in turn enables him or her use the access to the lower security host to exploit the access privileges granted to that host to attack a higher security host with better probability of success.

In a traditional scenario, the attacker would target a low-security Web server on a lower security part of the network by first scanning an organization's externally accessible IP addresses to gain access to a DMZ or internal network, and then use that access to be able to attack a database storing credit card numbers or other sensitive information. This is similar to attacking shared libraries in large programs to be able to inject malicious code into the overall program.

Pivot attacks in a Web 2.0 world are similar, but potentially much more dangerous, as Jeremiah Grossman points out. In this case, a Web 2.0 site will host content from a compromised external website that is programmed to serve up malicious code (JavaScript, for example), eventually attacking the victim's Web browser. Depending on how many affiliate levels there are in a network, the attack could come from a site many links away from the originating website. These attacks can install malware, manipulate cookies, or cause the various other amounts of damage outlined on Grossman's blog.

To prevent these types of attacks, minimize the amount of external party content on your website, vet the content prior to allowing it to be served up, automatically follow all of the links on your website and scan them for malicious code.

One of the protections that Jeremiah Grossman points out is sandboxing. Sandboxing would limit code to only access objects or data in its sandbox and not access everything that Web browser could access. This would mean that the malicious code potentially would only be able to access the objects it had access to in its sandbox, rather than the more broad access that a Web browser might have access to.

This was last published in March 2011

Dig Deeper on Web Server Threats and Countermeasures