Common Vulnerability Enumeration
These vulnerabilities are tracked as Common Vulnerabilities and Exposures (originally called Common Vulnerability Enumeration—CVEs). The score, also known as the CVE number, determines the severity of the vulnerability. These CVEs are identified and tracked by an organization in the U.S. As a system administrator, you can use these CVEs to implement updates to operating systems and software. This helps you understand how important the issue is and what the urgency level is.
Lifetime support?
With operating systems, a distinction is often made between major versions and minor versions. For example, my Mac laptop is currently running Mac OS Golden Gate (version 27) and subversion 2—in short, 27.2. Over time, the vendor’s support expires (Mac OS X release cycle). After that, you’re expected to perform a major upgrade to version 28 (which isn’t available yet, but just as an example). With Red Hat Linux, for instance, there’s 5 years of full support. After that, you can choose to start using a new major version, but you can also stay with the original version and purchase “Extended Support.” During these 5 years of extended support, no new minor versions will be released.
If you still don’t want to upgrade to the new major version, you can purchase 4 years of Extended Life Cycle support. Again, only security updates will be available under this plan. Finally, you can also purchase 4 years of Long Life support. In total, this allows you to benefit from support for 18 years. This is important for businesses, and there’s a simple reason for that.
The maintenance cost
I previously wrote about the business logic of an IT department or IT organization. My conclusion is that maintenance—maintaining systems and applications and keeping them operational—is actually the department’s or organization’s primary task. The risk here is that maintenance requires so much effort that the execution of IT projects suffers as a result. That’s a problem, because projects—and their results—actually accelerate the organization’s progress. And that’s of vital importance. This leads to the idea that maintenance must therefore be carried out as efficiently as possible. This, of course, ties into the discussion surrounding updates and upgrades. Much of the maintenance effort is actually tied up in updates and upgrades. It’s no coincidence, then, that Red Hat offers 18 years of support. The idea is that this saves you on maintenance work.
Ubuntu release cycle
Ubuntu Linux takes a slightly different approach. A minor version is released every six months, always in April and November. This year’s versions are 26.04 (which I’m running myself) and version 26.10 will be released soon. These versions receive six months of support. Every two years, a version is released with extended support. This is called a “Long-Term Support” (LTS) version. The latest LTS version is 24.04. If you don’t want to upgrade constantly, you should choose an LTS version. This is particularly important for business applications. Incidentally, during this support period, functional and security updates are available. With Linux distributions (such asRed Hat, Ubuntu, Debian, and SUSE ), these updates apply not only to the operating system but also to all applications and software provided through a repository (or a type of app store). This means you get much more software than commercial operating systems offer. It should be clear that other distributions also have a release cycle and support offerings.
By the way, I’ve been upgrading (major versions) my Ubuntu installations for years without any problems. I know this is also possible with Debian. Of course, there’s no guarantee it will work. It often depends on how complex the system is and the software installed on it. In that sense, it’s somewhat similar to the concept of“Rolling Releases,”where there are no fixed version milestones, but you install new versions as they become available. Examples of Linux distributions that follow this strategy include Gentoo Linux and Arch Linux.
Not all updates are the same
Many system administrators distinguish between functional updates and security updates (both are also referred to as patches). Functional updates introduce new features and, for example, fix functional bugs. Security updates specifically address security vulnerabilities that a system may have. It stands to reason that you’ll want to apply security updates as soon as possible (depending on the CVE number—urgency—see later in this article). So if you’re looking to set priorities, this is where you can do so. You could wait on functional updates, provided you aren’t experiencing the functional errors they fix.
Installing updates often brings new problems, and that’s the last thing you need. Of course, a new version of software always contains zero-day vulnerabilities that are discovered later on. That’s to be expected. But it’s very frustrating when existing (in-house) software fails because of updates. And that can definitely happen. And when existing software fails, it creates a lot of extra work—something we certainly don’t want to deal with in maintenance.
Strategy Update
I’ve come across IT organizations that, for this reason, always wait a while before implementing updates and patches. They always want to stay one minor version behind the market. They call this the N-1 strategy. This minimizes the risk of software failure with new versions. But at the same time, I think to myself, there are always zero-day vulnerabilities—always, by definition.
Infrastructure design also matters
Nowadays, security isn’t determined solely by updates, upgrades, and patches. The overall design of the infrastructure also plays a significant role in determining how secure the application ultimately is. Restrictions implemented through firewalls, proxies, and reverse proxies—as well as the use of Security Groups (AWS) in the cloud—determine the overall security of the system. Many of these measures are implemented during the design and implementation phases. The management burden stems primarily from repetitive tasks, updates, and upgrades/patches.
Automate Anything
As mentioned earlier, maintenance is primarily about efficiency. The adage “Automate as much as possible” fits well with this. Using Puppet, SaltStack, or Ansible, you can describe a system’s configuration and thus quickly rebuild it. In my opinion, automation plays a crucial role in keeping the management burden low. That’s the reason to adopt Infrastructure as Code (IaC).
CIA and EPSS
One strategy is to distinguish between different software stacks. For operating systems in a business setting, it’s a good idea to choose the LTS version—regardless of whether the software is open source or commercial. This allows the organization to operate for the longest possible time without having to upgrade, while still receiving security updates (minimal management effort). In addition, the Confidentiality, Integrity, and Availability (CIA) score is important. You can also look at the“Exploit Prediction Scoring System”(EPSS) score. This is an estimate of the likelihood that an exploit can actually be exploited.
The idea is that the higher the score, the sooner it deserves attention. For example, you can create groups: critical, high, medium, and low. Each group then has a different maximum timeframe within which the issue must be resolved. This can then be scheduled. Furthermore, publicly available systems are always a priority. And here, too, the more we automate, the better. A higher degree of automation means greater efficiency.
A curse or a blessing?
It is clear from the above that updates, patches, and upgrades are neither a curse nor a blessing. They are, however, necessary. Given the efficiency “requirement” of maintenance, it is important to automate this process as intelligently as possible. A well-designed strategy helps in this regard.
Current Events
A large number of disruptions were reported this week, especially in the United States. Microsoft, for example, was affected. Reports stated that this was not a hack, but a technical glitch. Earlier, supermarkets were affected by cooling systems that started heating instead of cooling. That incident was indeed a hack. But it wasn’t limited to these companies; Apple and Google were also affected. Rob Braxman explains it in detail. His conclusion is that AI is being used to find software vulnerabilities, such as zero-day exploits. He argues that we must therefore also use AI to counter this. Finding exploits is now happening so quickly that we cannot afford not to use AI. An interesting detail in his story is that the major vendors (Apple, Google, and Microsoft) get early access to the discovered exploits, but the Linux (community) does not. Whereas the vendors can fix problems quietly, the Linux community can only do so publicly. How does this relate to the topic of this article? Rob Braxman argues that the fixes are made so quickly that they introduce new bugs.
Incidentally, he also argues that the U.S. is leading the way in this development and that Europe(?) will follow. I wonder if we’re sufficiently prepared for this.