Viewing the future of Linux security with Crazy Penguin series

Article title: The Crazy Penguin series looks at the future of Linux security. Linux is a technology channel of the IT lab in China. Includes basic categories such as desktop applications, Linux system management, kernel research, embedded systems, and open source.

Don't think that your Linux system is safer just because Linux is safer than other operating systems. how can developers and publishers help system administrators in the future?

Do you know I have been writing this column for five years? What happened in the past five years? we can see that the linux competitor has accepted it, and its linux can already be used as a desktop platform.

In the field of linux security, there are also notable Developments. the linux firewall is now very mature, and a large number of embedded firewall devices are based on it, linux is also widely used for devices unrelated to security. linux supports surprisingly large numbers of security tools, making it the favorite system of security auditors and security consultants. In addition, linux has formed a role-based access control operating system, the most famous is NSA's SELinux.

But how can we raise the future of linux security? I have written a lot about the current situation and the past of linux security, but I have not written about the future. This month, I will summarize what I think of as linux security and where it should be.

　　Are there any errors?

Recently, it has been shown that a large number of people are not much more secure than Microsoft's windows operating system. before the discussion, let's explain this point first. first, personally, I think linux is safer than windows. I have already repeated this article in this column, in linux, users have easier control over their systems than in windows.

The problem is that linux users, like windows users, tend to concentrate their efforts on letting the system do what they want the system to do. they trust the system's built-in or default security settings too much, when unavoidable software bugs occur, the impacts of those bugs tend to be wider, and more work is needed than prevention.

For example, if we run BINDv9 to provide the name resolution service, it will take some work and research to make it work. more efforts are required to put it in a chroot environment, chroot can run the named process in a subset of the file system. Therefore, when a BIND vulnerability is discovered, most BIND users may have experienced the pain of not putting it in the chroot environment, if Microsoft's name resolution service is running-it does not have as many security features as the BIND, the pain index may be the same.

All of this is to simply tell you that most of the linux security features and functions do not benefit linux users. as a result, at least according to what my friends said, professional penetration testing should be conducted, breaking your general Redhat Enterprise Edition is not more difficult than the general windows2003 system.

This is unfortunate and surprising. Its code is completely transparent, and linux still has similar software bugs. In general, the number and frequency of windows are almost the same. Like windows, linux is a lot of complex code developed by hundreds of people. the more code there are, the more bugs there may be, right?

In a recent interview with SearchSecurity.com, I talked about a security innovation article on the guidance of the Microsoft Research Fund. The results show that windows is safer than linux, the conclusion is mainly based on common security bugs and the average release time of patches. I believe that I have correctly commented on the research results without considering other linux security advantages, such as customization capabilities and software package options, in other words, I feel that this study is more about comparing the default installation scenarios without considering the factors that each operating system uses for security reinforcement.

But I think more about it. I am more worried that the security risks of a platform may not be counted, unless the platform running most systems actually triggers hidden risks, strictly speaking, this is not an end-user action, and I will not blame the system administrator because I will elaborate on it later, I think linux developers and publishers must continue to want to make the security features more common, transparent, and easier to configure and use, by the way, because I am comparing linux and windows, to be fair, I should point out that windows also has many security features, but users rarely use them.

Good. by default, both linux and windwos are insecure, and both software bugs and security patches are comparable.

Both operating systems use simple access control mode for security settings. In this mode, a super user account is root in linux, in windows, Administrator-has the right to control all systems, including files of other users. in these two operating systems, group members can be used to create access of different levels. for example, root can perform multiple authorizations. In fact, in most systems, you have to log on as a privileged user or temporarily become the user in order to accomplish important tasks.

As a result, any process running by a privileged user can have full control over the linux or windwos system. However, I log on to and configure important background processes as a non-privileged user, the bug of these background processes does not affect the entire system. However, bugs in other software may allow it to upgrade its permissions from a non-root process. for example, if you have obtained a web server running Apache, one day, an attacker controlled the attack program and attacked an unpatched Apache buffer overflow vulnerability. as a result, the attacker obtained a shell session on your server, the attacker is running as www, because Apache is running as www. if there is an unpatched kernel vulnerability in the system, the local permission will be improved.

The system administrator may already know the vulnerability, but the patch has not yet been released. after all, it is strictly a local vulnerability. no one except you has shell permission on the system, who wants to have to restart after patching the kernel? But now a remote attacker has a local shell access permission. if he successfully exploits this kernel vulnerability, he will be the root user. this is a common intrusion scenario, but you don't have to worry about this after using the root-takes-all security module.

This is the current situation of linux Security. to protect linux, we need to spend a considerable amount of effort to utilize complex security features. these complex features are often not enabled by default, keep all patches up-to-date. In a good company, most of us use modern operating systems with the same limitations and challenges.

[1] [2] [3] Next page