This book explores the Linux platform and various Linux-based operating systems – in particular, how Linux can work well for networking services. We'll start by discussing some of the history of the operating system before looking at its basic configuration and troubleshooting. From there, we'll work through building various network-related services on Linux that you may commonly see in most organizations. As we progress, we'll build real services on real hosts, with an emphasis on securing and troubleshooting each service as we go. By the time we're done, you should be familiar enough with each of these services to start implementing some or all of them in your own organization. As they say, every journey begins with a single step, so let's take that step and start with a general discussion of the Linux platform.

In this chapter, we'll start our journey by exploring Linux as a family of operating systems. They're all related, but each is unique in its own way, with different strengths and features.

We'll cover the following topics:

• Why Linux is a good fit for a networking team
• Mainstream data center Linux
• Specialty Linux distributions
• Virtualization
• Picking a Linux distribution for your organization

In this book, we'll explore how to support and troubleshoot your network using Linux and Linux-based tools, as well as how to securely deploy common networking infrastructure on Linux platforms.

Why would you want to use Linux for these purposes? To begin with, the architecture, history, and culture of Linux steers administrators toward scripting and automating processes. While carrying this to extremes can get people into funny situations, scripting routine tasks can be a real time-saver.

In fact, scripting non-routine tasks, such as something that needs doing once per year, can be a lifesaver as well – it means that administrators don't need to relearn how to do that thing they did 12 months ago.

Scripting routine tasks is an even bigger win. Over many years, Windows administrators have learned that doing one task hundreds of times in a Graphical User Interface (GUI) guarantees that we misclick at least a few times. Scripting tasks like that, on the other hand, guarantees consistent results. Not only that, but over a network, where administrators routinely perform operations for hundreds or thousands of stations, scripting is often the only way to accomplish tasks at larger scales.

Another reason that network administrators prefer Linux platforms is that Linux (and before that, Unix) has been around since there were networks to be a part of. On the server side, Linux (or Unix) services are what defined those services, where the matching Windows services are copies that have mostly grown to feature parity over time.

On the workstation side, if you need a tool to administer or diagnose something on your network, it's probably already installed. If the tool that you seek isn't installed, it's a one-line command to get it installed and running, along with any other tools, libraries, or dependencies required. And adding that tool does not require a license fee – both Linux and any tools installed on Linux are (almost without exception) free and open source.

Lastly, on both the server and desktop side, historically, Linux has been free. Even now, when for-profit companies have license fees for some of the main supported distributions (for instance, Red Hat and SUSE), those companies offer free versions of those distributions. Red Hat offers Fedora Linux and CentOS, both of which are free and, to one extent or another, act as test-bed versions for new features in Red Hat Enterprise Linux. openSUSE (free) and SUSE Linux (chargeable) are also very similar, with the SUSE distribution being more rigorously tested and seeing a more regular cadence for version upgrades. The enterprise versions are typically term-licensed, with that license granting the customer access to technical support and, in many cases, OS updates.

Many companies do opt for the licensed enterprise-ready versions of the OS, but many other companies choose to build their infrastructures on free versions of OpenSUSE, CentOS, or Ubuntu. The availability of free versions of Linux means that many organizations can operate with substantially lower IT costs, which has very much influenced where we have gone as an industry.

Why is Linux important?

Over the years, one of the jokes in the information technology community is that next year was always going to be the year of the Linux desktop – where we'd all stop paying license fees for desktops and business applications, and everything would be free and open source.

Instead, what has happened is that Linux has been making steady inroads into the server and infrastructure side of many environments.

Linux has become a mainstay in most data centers, even if those organizations think they are a Windows-only environment. Many infrastructure components run Linux under the covers, with a nice web frontend to turn it into a vendor solution. If you have a Storage Area Network (SAN), it likely runs Linux, as do your load balancers, access points, and wireless controllers. Many routers and switches run Linux, as do pretty much all the new software-defined networking solutions.

Almost without fail, information security products are based on Linux. Traditional firewalls and next-generation firewalls, Intrusion Detection and Prevention Systems (IDS/IPS), Security Information and Event Management (SIEM) systems, and logging servers – Linux, Linux, Linux!

Why is Linux so pervasive? There are many reasons:

• It is a mature operating system.
• It has an integrated patching and updating system.
• The basic features are simple to configure. The more complex features on the operating system can be more difficult to configure than on Windows though. Look ahead to our chapter on DNS or DHCP for more information.
• On the other hand, many features that might be for sale products in a Windows environment are free to install on Linux.
• Since Linux is almost entirely file-based, it's fairly easy to keep it to a known baseline if you are a vendor who's basing their product on Linux.
• You can build just about anything on top of Linux, given the right mix of (free and open source) packages, some scripting, and maybe some custom coding.
• If you pick the right distribution, the OS itself is free, which is a great motivator for a vendor trying to maximize profit or a customer trying to reduce their costs.

If the new Infrastructure as Code movement is what draws you, then you'll find that pretty much every coding language is represented in Linux and is seeing active development – from new languages such as Go and Rust, all the way back to Fortran and Cobol. Even PowerShell and .NET, which grew out of Windows, are completely supported on Linux. Most infrastructure orchestration engines (for instance, Ansible, Puppet, and Terraform) started on and supported Linux first.

On the cloud side of today's IT infrastructure, the fact that Linux is free has seen the cloud service providers push their clients toward that end of the spectrum almost from the start. If you've subscribed to any cloud service that is described as serverless or as a Service, behind the scenes, it's likely that that solution is almost all Linux.

Finally, now that we've seen the server and infrastructure side of IT move toward Linux, we should note that today's cell phones are steadily becoming the largest desktop platform in today's computing reality. In today's world, cell phones are generally either iOS- or Android-based, both of which are (you guessed it) Unix/Linux-based! So, the year of the Linux desktop has snuck upon us by changing the definition of desktop.

All of this makes Linux very important to today's networking or IT professionals. This book focuses on using Linux both as a desktop toolbox for the networking professional, as well as securely configuring and delivering various network services on a Linux platform.

The history of Linux

To understand the origins of Linux, we must discuss the origins of Unix. Unix was developed in the late 1960s and early 1970s at Bell Labs. Dennis Ritchie and Ken Thompson were Unix's main developers. The name Unix was actually a pun based on the name Multics, an earlier operating system that inspired many of Unix's features.

In 1983, Richard Stallman and the Free Software Foundation started the GNU (a recursive acronym – GNU's Not Unix) project, which aspired to create a Unix-like operating system available to all for free. Out of this effort came the GNU Hurd kernel, which most would consider the precursor to today's Linux versions (the SFS would prefer we called them all GNU/Linux).

In 1992, Linus Torvalds released Linux, the first fully realized GNU kernel. It's important to note that mainstream Linux is normally considered to be a kernel that can be used to create an operating system, rather than an operating system on its own. Linux is still maintained with Linus Torvalds as the lead developer, but today, there is a much larger team of individuals and corporations acting as contributors. So, while technically Linux only refers to the kernel, in the industry, Linux generally refers to any of the operating systems that are built upon that kernel.

Since the 1970s, hundreds of separate flavors of Linux have been released. Each of these is commonly called a distribution (or distro, for short). These are each based on the Linux kernel of the day, along with an installation infrastructure and a repository system for the OS and for updates. Most are unique in some way, either in the mix of base packages or the focus of the distro – some might be small in size to fit on smaller hardware platforms, some might focus on security, some might be intended as a general-purpose enterprise workhorse operating system, and so on.

Some distros have been "mainstream" for a period of time, and some have waned in popularity as time has gone by. The thing they all share is the Linux kernel, which they have each built upon to create their own distribution. Many distros have based their operating system on another distro, customizing that enough to justify calling their implementation a new distribution. This trend has given us the idea of a "Linux family tree" – where dozens of distributions can grow from a common "root." This is explored on the DistroWatch website at https://distrowatch.com/dwres.php?resource=family-tree.

An alternative to Linux, especially in the Intel/AMD/ARM hardware space, is Berkeley Software Distribution (BSD) Unix. BSD Unix is a descendent of the original Bell Labs Unix; it is not based on Linux at all. However, BSD and many of its derivatives are still free and share many characteristics (and a fair amount of code) with Linux.

To this day, the emphasis of both Linux and BSD Unix is that both are freely available operating systems. While commercial versions and derivatives are certainly available, almost all those commercial versions have matching free versions.

In this section, we looked at both the history and importance of Linux in the computing space. We understood how Linux emerged and how it found popularity in certain sections of the computing landscape. Now, we'll start looking at the different versions of Linux that are available to us. This will help us build on the information we need to make choices regarding which distro to use later in this chapter.

As we've discussed, Linux is not a monolithic "thing," but rather a varied or even splintered ecosystem of different distributions. Each Linux distribution is based on the same GNU/Linux kernel, but they are packaged into groups with different goals and philosophies, making for a wide variety of choices when an organization wants to start standardizing on their server and workstation platforms.

The main distributions that we commonly see in modern data centers are Red Hat, SUSE, and Ubuntu, with FreeBSD Unix being another alternative (albeit much less popular now than in the past). This is not to say that other distributions don't crop up on desktops or data centers, but these are the ones you'll see most often. These all have both desktop and server versions – the server versions often being more "stripped down," with their office productivity, media tools, and, often, the GUI removed.

Red Hat

Red Hat has recently been acquired by IBM (in 2019), but still maintains Fedora as one of its main projects. Fedora has both server and desktop versions, and remains freely available. The commercial version of Fedora is Red Hat Enterprise Linux (RHEL). RHEL is commercially licensed and has a formal support channel.

CentOS started as a free, community-supported version of Linux that was functionally compatible with the Red Hat Enterprise version. This made it very popular for server implementations in many organizations. In January 2014, Red Hat pulled CentOS into its fold, becoming a formal sponsor of the distro. In late 2020, it was announced that CentOS would no longer be maintained as a RHEL-compatible distribution but would rather "fit" somewhere between Fedora and RHEL – not so new as to be "bleeding edge," but not as stable as RHEL either. As part of this change, CentOS was renamed CentOS Stream.

Finally, Fedora is the distro that has the latest features and code, where new features get tried and tested. The CentOS Stream distro is more stable but is still "upstream" of RHEL. RHEL is a stable, fully tested operating system with formal support offerings.

Oracle/Scientific Linux

Oracle/Scientific Linux is also seen in many data centers (and in Oracle's cloud offerings). Oracle Linux is based on Red Hat, and they advertise their product as being fully compatible with RHEL. Oracle Linux is free to download and use, but support from Oracle is subscription-based.

SUSE

OpenSUSE is the community distribution that SUSE Linux is based on, similar to how RedHat Enterprise Linux is based on Fedora.

SUSE Linux Enterprise Server (commonly called SLES) was, in the early days of Linux, the mainly European competitor for the US-based Red Hat distribution. Those days are in the past, however, and SUSE Linux is (almost) as likely to be found in Indiana as it is in Italy in modern data centers.

Similar to the relationship between RedHat and CentOS, SUSE maintains both a desktop and a server version. In addition, they also maintain a "high-performance" version of the OS, which comes with optimizations and tools pre-installed for parallel computing. OpenSUSE occupies an "upstream" position to SLES, where changes can be introduced in a distro that is somewhat more "forgiving" to changes that might not always work out the first time. The OpenSUSE Tumbleweed distro has the newest features and versions, where as OpenSUSE Leap is closer in versioning and stability to the SLE versions of the operating system. It is no accident that this model is similar to the RedHat family of distros.

Ubuntu

Ubuntu Linux is maintained by Canonical and is free to download, with no separate commercial or "upstream" options. It is based on Debian and has a unique release cycle. New versions of both the server and desktop versions are released every 6 months. A Long-Term Support (LTS) version is released every 2 years, with support for LTS versions of both the server and desktop running for 5 years from the release date. As with the other larger players, support is subscription-based, though free support from the community is a viable option as well.

As you would expect, the server version of Ubuntu is focused more on the core OS, network, and data center services. The GUI is often de-selected during the installation of the server version. The desktop version, however, has several packages installed for office productivity, media creation, and conversion, as well as some simple games.

BSD/FreeBSD/OpenBSD

As we mentioned previously, the BSD "tree" of the family is derived from Unix rather than from the Linux kernel, but there is lots of shared code, especially once you look at the packages that aren't part of the kernel.

FreeBSD and OpenBSD were historically viewed as "more secure" than the earlier versions of Linux. Because of this, many firewalls and network appliances were built based on the BSD OS family, and remain on this OS to this day. One of the more "visible" BSD variants is Apple's commercial operating system OS X (now macOS). This is based on Darwin, which is, in turn, a fork of BSD.

As time marched on, however, Linux has grown to have most of the same security capabilities as BSD, until BSD perhaps had the more secure default setting than most Linux alternatives.

Linux now has security modules available that significantly increase its security posture. SELinux and AppArmor are the two main options that are available. SELinux grew out of the Red Hat distros and is fully implemented for SUSE, Debian, and Ubuntu as well. AppArmor is typically viewed as a simpler-to-implement option, with many (but not all) of the same features. AppArmor is available on Ubuntu, SUSE, and most other distros (with the notable exception of RHEL). Both options take a policy-based approach to significantly increase the overall security posture of the OS they are installed on.

With the evolution of Linux to be more security focused, in particular with SELinux or AppArmor available (and recommended) for most modern Linux distributions, the "more secure" argument of BSD versus Linux is now mainly a historic perception rather than fact.

Aside from the mainstream Linux distributions, there are several distros that have been purpose-built for a specific set of requirements. They are all built on a more mainstream distro but are tailored to fit a specific set of needs. We'll describe a few here that you are most likely to see or use as a network professional.

Most commercial Network-attached Storage (NAS) and SAN providers are based on Linux or BSD. The front runner on open source NAS/SAN services, at the time of writing, seems to be TrueNAS (formerly FreeNAS) and XigmaNAS (formerly NAS4Free). Both have free and commercial offerings.

Open source firewalls

Networking and security companies offer a wide variety of firewall appliances, most of which are based on Linux or BSD. Many companies do offer free firewalls, some of the more popular being pfSense (free versions and pre-built hardware solutions available), OPNsense (freely available, with donations), and Untangle (which also has a commercial version). Smoothwall is another alternative, with both free and commercial versions available.

In this book, we'll explore using the on-board firewall in Linux to secure individual servers, or to secure a network perimeter.

Kali Linux

Descended from BackTrack, and KNOPPIX before that, Kali Linux is a distribution based on Debian that is focused on information security. The underlying goal of this distribution is to collect as many useful penetration testing and ethical hacking tools as possible on one platform, and then ensure that they all work without interfering with each other. The newer versions of the distribution have focused on maintaining this tool interoperability as the OS and tools get updated (using the apt toolset).

SIFT

SIFT is a distribution authored by the forensics team at the SANS institute, focused on digital forensics and incident response tools and investigations. Similar to Kali, the goal of SIFT is to be a "one-stop shop" for free/open source tools in one field – Digital Forensics and Incident Response (DFIR). Historically, this was a distribution based on Ubuntu, but in recent years, this has changed – SIFT is now also distributed as a script that installs the tools on Ubuntu desktop or Windows Services for Linux (which is Ubuntu-based).

Security Onion

Security Onion is also similar to Kali Linux in that it contains several information security tools, but its focus is more from the defender's point of view. This distribution is centered on threat hunting, network security monitoring, and log management. Some of the tools in this distribution include Suricata, Zeek, and Wazuh, just to name a few.

Virtualization has played a major role in the adoption of Linux and the ability to work with multiple distributions at once. With a local hypervisor, a network professional can run dozens of different "machines" on their laptop or desktop computers. While VMware was the pioneer in this space (desktop and dedicated virtualization), they have since been joined by Xen, KVM, VirtualBox, and QEMU, just to name a few. While the VMware products are all commercial products (except for VMware Player), the other solutions listed are, at the time of writing, still free. VMware's flagship hypervisor, ESXi, is also available for free as a standalone product.

Linux and cloud computing

The increasing stability of Linux and the fact that virtualization is now mainstream has, in many ways, made our modern-day cloud ecosystems possible. Add to this the increasing capabilities of automation in deploying and maintaining backend infrastructure and the sophistication available to the developers of web applications and Application Programming Interfaces (APIs), and what we get is the cloud infrastructures of today. Some of the key features of this are as follows:

• A multi-tenant infrastructure, where each customer maintains their own instances (virtual servers and virtual data centers) in the cloud.
• Granular costing either by month or, more commonly, by resources used over time.
• Reliability that it is as good or better than many modern data centers (though recent outages have shown what happens when we put too many eggs in the same basket).
• APIs that make automating your infrastructure relatively easy, so much so that for many companies, provisioning and maintaining their infrastructure has become a coding activity (often called Infrastructure as Code).
• These APIs make it possible to scale up (or down) on capacity as needed, whether that is storage, computing, memory, session counts, or all four.

Cloud services are in business for a profit, though – any company that has decided to "forklift" their data center as is to a cloud service has likely found that all those small charges add up over time, eventually reaching or surpassing the costs of their on-premises data center. It's still often attractive on the dollars side, as those dollars are spent on operational expenses that can be directly attributed more easily than the on-premises capital expenditure model (commonly called Cap-Ex versus Op-Ex models).

As you can see, moving a data center to a cloud service does bring lots of benefits to an organization that likely wouldn't have the option to in the on-premises model. This only becomes more apparent as more cloud-only features are utilized.

In many ways, which distribution you select for your data center is not important – the main distributions all have similar functions, often have identical components, and often have similar vendor or community support options. However, because of the differences between these distros, what is important is that one distribution (or a set of similar distros) is selected.

The desired outcome is that your organization standardizes one distribution that your team can develop their expertise with. This also means that you can work with the same escalation team for more advanced support and troubleshooting, whether that is a consulting organization, a paid vendor support team, or a group of like-minded individuals on various internet forums. Many organizations purchase support contracts with one of "the big three" (Red Hat, SUSE, or Canonical, depending on their distribution).

Where you don't want to be is in the situation I've seen a few clients end up in. Having hired a person who is eager to learn, a year later, they found that each of the servers they built that year were on a different Linux distribution, each built slightly differently. This is a short road to your infrastructure becoming the proverbial "science experiment" that never ends!

Contrast this with another client – their first server was a SUSE Linux for SAP, which is, as the name suggests, a SUSE Linux server, packaged with the SAP application that the client purchased (SAP HANA). As their Linux footprint grew with more services, they stuck with the SUSE platform, but went with the "real" SLES distribution. This kept them on a single operating system and, equally important for them, a single support license with SUSE. They were able to focus their training and expertise on SUSE. Another key benefit for them was that as they added more servers, they were able to apply a single "stream" of updates and patches with a phased approach. In each patch cycle, less critical servers got patched first, leaving the core business application servers to be patched a few days later, after their testing was complete.

The main advice in picking a distribution is to stick to one of the larger distributions. If people on your team have strong feelings about one of these, then definitely take that into consideration. You will likely want to stay fairly close to one of the mainstream distributions so that you can use it within your organization, something that is regularly maintained and has a paid subscription model available for support – even if you don't feel you need paid support today, that may not always be the case.

Now that we've discussed the history of Linux, along with several of the main distributions, I hope you are in a better position to appreciate the history and the central importance of the operating systems in our society. In particular, I hope that you have some good criteria to help you choose a distro for your infrastructure.

In this book, we'll choose Ubuntu as our distribution. It's a free distribution, which, in its LTS version, has an OS that we can depend on being supported as you work through the various scenarios, builds, and examples that we'll discuss. It's also the distribution that is native to Windows (in Windows services for Linux). This makes it an easy distro to become familiar with, even if you don't have server or workstation hardware to spare or even a virtualization platform to test with.

In the next chapter, we'll discuss getting your Linux server or workstation on the network. We'll illustrate working with the local interfaces and adding IP addresses, subnet masks, and any routes required to get your Linux host working in a new or existing network.

• Red Hat Linux: https://www.redhat.com/en
• Fedora: https://getfedora.org/
• CentOS: https://www.centos.org/
• SUSE Linux: https://www.suse.com/
• OpenSUSE: https://www.opensuse.org/
• Ubuntu Linux: https://ubuntu.com/
• Windows Subsystem for Linux: https://docs.microsoft.com/en-us/

  https://docs.microsoft.com/en-us/windows/wsl/about

• FreeBSD Unix: https://www.freebsd.org/
• OpenBSD Unix: https://www.openbsd.org/
• Linux/BSD differences: https://www.howtogeek.com/190773/htg-explains-whats-the-difference-between-linux-and-bsd/
• TrueNAS: https://www.truenas.com/
• XigmaNAS: https://www.xigmanas.com/
• pfSense: https://www.pfsense.org/
• OPNsense: https://opnsense.org/
• Untangle: https://www.untangle.com/untangle
• Kali Linux: https://www.kali.org/
• SIFT: https://digital-forensics.sans.org/community/downloads; https://www.sans.org/webcasts/started-sift-workstation-106375
• Security Onion: https://securityonionsolutions.com/software
• Kali Linux: https://www.kali.org/