Breaking In

In June 2009, a virtualization product aimed at web servers was found to have a few security vulnerabilities. The end result was that about 100,000 web sites got hacked and deleted at a number of different providers. (It's not clear how many were recovered.) Also in June, the security-related website astalavista.com was hacked, and a variety of files and databases, as well as the remote backups, were deleted. These were only the "big" hacks that were newsworthy, the actual number of websites and servers compromised is much higher.

The techniques for network attacks keep evolving. In this article, I take a look at some favorite strategies for the latest generation of intruders.

A Quick Legal Disclaimer

Please note that engaging in the kinds of activities described in this article can potentially get you into trouble, ranging from a stern talking to by your network administrator to a less-than-enjoyable, all-expenses-paid vacation courtesy of whichever law enforcement agency you manage to annoy the most. So why am I writing this? If you want to build and maintain secure systems, you need to understand how to make them fail. If you want to buy a good lock, you either have to buy a bunch of locks and learn how they work or find someone who has [1]. My advice is to get a cheap quad-core machine with lots of RAM, put VirtualBox or VMware on it, and build systems and networks you can attack without disturbing anyone else.

A Brief History

Life used to be pretty simple. You had a server, and on it you ran a couple of services (mail, file, DNS, etc.). If users wanted an application, you installed it on their machines. If users wanted to edit or upload content remotely to the web, you gave them FTP access. Email was just text, PDF files didn't include JavaScript, and image files were just image files – they weren't executable content. To secure your network, you simply kept things up to date, firewalled access, and ran as many services as possible without root access.

On Brute Force Attacks

Some automated tools simply hammer away, attempting a variety of common exploits against any server they can connect to, giving up speed and sophistication for brute force. This often works because of the sheer number of web servers and applications and, more importantly, because of the number of out-of-date applications with well-known security flaws (witness Adobe taking several weeks to months to fix various vulnerabilities in their Reader product). Some studies put the percentage of abandoned web logs at 95%, and, if no one is updating them with content, the chances are that no one is updating them for security fixes [2].