Security in Linux Linux, like any other computing platform, is constantly changing. There are a few major focus points for new and upgraded platforms, one of which is how user friendly it is. User friendliness goes beyond the ability to simply point and click, it also goes behind the lines deep into the inner workings of the system. Security is one of the most important functions of any operating system, very commonly overlooked and taken for granted. A system administrator can configure tables that are provided by the Linux kernel firewall in a program called iptables. Iptables has the ability to redirect, modify or stop packets of data all based on the state of a connection at any given time. There are many different tables that can be defined and each table contains built in chains or user defined chains. Every chain is essentially a list of rules that matches a set of packets and it specifies what to do with a packet that matches the rules. For the casual user it is best to use the predefined rules, they are often more than adequate. In an enterprise situation the administrator would likely want to define additional rules in order to best suit the business needs. Before iptables Linux mainly used ipchains as a firewall package. Iptables is an improvement on ipchains because it monitors the state of connections. Iptables can use the state of the connection as opposed to ipchains using the source destination and content only, to redirect, modify or drop a packet. At least one chain is used for every packet leaving or entering the computer on the network. The initial chain used is determined by the origin of the packet.

Another powerful and important tool in the security of Linux is the chroot jail. Chroot made its first appearance in 1979 in the 7th version of Unix and was added to BSD in 1982 (Wikipedia, 2011). Chroot jail has many uses in the Linux world. It can be used to prevent a single user from going to certain parts of the directory, it can be used in virtualization or it can be used for specific processes. For singer users, even one with limited permissions, without a chroot jail the user could navigate to upper level directories. Even if an administrator trusts their users (and they never should), without a chroot jail the information is there for anyone who can hack the users account. The specific user may not have permissions to edit other user’s files but he may be able to see them and that can be dangerous on its own. Testing is another great argument for chroot jails, if a person desires to test a new program it is wise to limit the programs access to system resources until it is certain that there are no malicious intents inserted into the program itself. In virtualization a chroot jail is also quite necessary. A person could have a complete operating system installed in a chroot jail and it would appear as though the chroot directory was the root of the system. This would also aid in testing and development because the guest system is completely contained within the jail and it wouldn’t be able to make changes to the host system. This would be a vital tool to utilize in system program compatibility testing, if anything undesirable happened then the user could simply revert to an original and unaltered virtual machine file and start fresh. A chroot jail is not specifically meant for intentional attacks but more as a means of preventing inadvertent system altering mistakes made by the casual user.

SELinux or Security Enhanced Linux is an implementation of flexible mandatory access controls that was first researched and implemented into the Linux kernel by researchers at the National Security Agency and the Secure Computing Corporation in 1992 and 1993. (Flux Advanced Security Kernel, 2000) As partially stated above, SELinux is the implementation of mandatory access controls in the Linux whereas standard Linux uses discretionary access controls. In a system using discretionary access user identity and user ownership dictates file and resource decisions. If a user running root privileges starts a process that has malicious intent then the process can in theory take control of any file or system resource and change it to better suit the malicious intent of the process or the creator of the process. In a system running SELinux the mandatory access controls can be administratively designed so as to very closely monitor processes and software with malicious intent. Security decisions in SELinux are not bases only on the user identity alone but also on what security information is available at the time that the process begins. Using SELinux also helps to provide control not only on programs and processes but on users and devices as well. The mandatory access controls in SELinux are only as good or safe as the administrator sets them up to be, it is wise to have well developed and implemented policies in place at all times and for all users. It is important to note that SELinux was not a kernel rewrite but can be seen more like a patch that provides framework for future security improvements and it uses the Linux Security Module. While only three of the multitude of security technologies have been discussed here, there are so many more exciting security tools available in Linux. Some are targeted at the advanced user while others can be utilized by just about anyone. The bottom line is that with all of the tools and information available, there is no reason to be left unprotected. Every computer that is connected to the internet will likely have information that someone, somewhere wants to get. Keep your guard up and be safe.