Modern NetOps Needs More from Network Engineers
Introduction
Network Engineering and what it takes to thrive in this new reality is undergoing a massive shift. We need to understand how we got here, and how we can adapt and thrive in this new environment.?
The bottom line is that since the world has become more software-centric than ever, it now takes more skills and knowledge to design, engineer, deploy, and operate a network than before.?
This software-centric perspective of networking is a main component of “Total Network Operations” (TNOps for short), a framework we can use to adapt to this new reality.
This post is the first of a two-part series, where we’ll discuss tech disruption, what it means for network engineers, and how you can acquire skills to become a Next Gen Network Engineer.?
(Part two in this series will dive deep into Total Network Operations as the framework of what modern NetOps needs from IT and Network Operations organizations, and how those orgs and their leaders can adapt to support Next Gen Network Engineers in the software-centric networking world.)
Disruption Happens
If you’re reading this, it isn't because you woke up one morning and accidentally found yourself in the “tech” part of an organization. Whether it was networking, security, cloud, or another area, it’s likely that part of what attracted you was the cool new tech things you could learn, and the interesting things you could do with that knowledge. When that happened – whether meaning to or not – you essentially signed up to ride a wave of disruption, and you probably have enjoyed that ride up until this point. So what needs to change?
The Reality of Disruption
Disruption is a part of tech and IT overall, and over the last 10-15 years we’ve seen radical disruption and progress in many areas of tech. Some examples are Software-Defined Networking (SDN), software-centric tools and processed applied to networking, cloud, containerization/micro-services, and AI. No one does their work in these areas the way they did 10 years ago.
This type of radical, once-in-a-generation change is now happening in NetOps. Along with that, the new tools and technologies that will enable that change will require individuals to upskill.
In Part Two, we’ll talk more about how organizations need to “upskill” as well. It won’t be just the responsibility for individuals to change, but this new reality will also require organizations to examine how they use and manage tools and tech via new processes, new workflows, and updated ways of thinking about productivity, culture, and ROI.?
Houston, We May Have a Problem
We need to address a very real problem and personal issue. You may not be excited about the disruption you’re encountering as a Network Engineer, and the new tools and technology that will make you successful. I get it. It took a long time to build up the knowledge, skills, and abilities you have now, and you likely got into Network Engineering precisely because these are the technologies and tools you were excited to learn in the first place. If you’re upset about everything that’s changing for network engineering today, I have a couple of key messages for you:
Software Eats the World, Enabled by Hardware
Let’s take a look at disruption in an adjacent and slightly overlapping area: the world of Software and Computing.
Marc Andreesen, famous co-founder of Netscape Communications and founding partner of VC firm Andreesen-Horowitz, wrote about innovation via software in an August 2011 blog post titled “Why Software Is Eating the World”. It’s a highly recommended read.?
The essence of the quote is that software provides the ability to innovate rapidly - much more so than hardware, which takes longer to design, build, and deploy. The distribution of hardware is bound by the speed of delivery by jet (and then some).?
Distribution of software is bound by the speed of light, and having a robust global network. That networking hardware is absolutely critical and it will always be true that reliable, fast, and distributed networking hardware will be required to facilitate the rapid distribution of software.
Coming full circle, it is the rapid disruption enabled by software that is causing the disruption in NetOps. The big idea here is that since software is fungible, when combined with nimble processes, rapid iteration in software development generates products, features, and solutions we can use in networking like never before. The logical result is that as software and software-centric processes gather steam and deliver improved outcomes, we should expect an impact from these increasingly popular processes on how we design, architect, engineer, and operate networks.
Disruption in Software and Computing
As I noted above, Networking is not the first tech discipline to go through this disruption. Take a quick look at some of the changes that Software and Computing have gone through:
A Telco Example: Switchboards to Voice Switches to Ethernet Switches
We also have examples of disruption closer to home in telecommunications. Remember old TV show references to telephone operators helping complete calls? We used to have manual (!!!) switchboard operators where the calling party didn’t dial a number, but rather spoke to an operator and asked to be connected to another party. The operator then made physical patch connections between the two parties. This article from History.com is a good read on the topic.
Initially, this labor-intensive task was carried out by boys, but the boys tended to misbehave and get distracted, so young women were recruited to meet the rapidly-growing demand for telephone services. Rules placed on telephone operators were strict, and they began to organize and push back on companies for less strict rules and higher wages.
Because of the rapid adoption of telephony and these price pressures, in the 1930s, we took the “operator" function in the telco network and we started to automate it. First, we used mechanical switches based on digits dialed to automatically connect calls. Then, we created electronic switches (You may have heard of a 5ESS, the end of the line in that era of telco voice switches).?
Of course, those switches needed special software to operate. Today, even voice calls are data in packets, and circuit switching has been largely displaced by packet switching (which is what routers and switches do). As modern network engineers, we participate in and benefit from this disruption.
What about the poor telephone operators? Some switchboard operators were able to learn how to operate switches. Over time, switchboard operators and more modern operators were displaced, through the displacement of their previously manual functions in many enterprises.?
In addition to automated call completion, we saw a progression of new technologies and tools, including Private Branch Exchanges (PBXs), Centrex Service to displace PBXs, Automated Call Directors (ACDs) to reduce/eliminate the need for a person to direct your call (which we all love today, of course), and eventually Caller ID (so you know who was calling) and VoiceMail (so a person didn’t need to write down a message). Caller ID has been supplanted by your local contacts and spam call identification, and VoiceMail still persists today, but more and more of us don’t want to listen to voicemails - we’d rather read the auto-generated transcript.
The market didn’t remain stagnant. It expanded and hundreds of thousands of new jobs were created in installation and maintenance of switches, development of software and hardware for the new switches (not just once, but for generations of software and hardware), and manufacturing for those switches. This also accelerated the development of new services, such as the development of advanced signaling protocols such as SS7.
So, changes in the industry and demand for telephone service drove the market to make telephone calls more efficient. The manual switchboard approach was a critical step in the development of telephony, but it was disrupted, and we barely even think of those tools, roles, and events today.
Mastering Disruption: The Next Gen Network Engineer
If you’re ready to engage in learning these new skills and tech, you might be thinking, “Ok, what do I need to learn, and where do I start?”
Great questions, glad you asked. Here is a greatly simplified learning roadmap that you can use to make progress on becoming a NGNE today.?
In future posts we’ll lay out clearly and completely what we believe makes for a fully-trained and enabled NGNE, and point you to resources that you can use to do so.
0. Networking Basics
This is table stakes. We are talking about networking here, so you can’t skip this. You still need this base layer for subnetting, spanning tree, and more. If you already know this stuff, skip ahead. If not …
The Cisco Certified Network Associate (CCNA) is one of the most common and broadly applicable entry-level networking certs, mostly because of the? install base of Cisco networking equipment. It’s also a very practical place to start.
The CCNA is not the only way to start.? There is a lot of swapability here, and you can swap out to other certs and networking environments if you already know what equipment you’ll be working on. Two other very good examples are the Juniper JNCIA-Junos and the Arista ACE L1 certifications.
I’m calling out specific certifications here because these are programs that I’m personally familiar with and that I have confidence in recommending. There are certainly other certification programs that can fit the bill here, and there are vendor-neutral networking certification programs. While they can be helpful, you will still need to figure out how to implement those concepts on a set of vendor equipment, so starting with a vendor-specific certification program can actually save you time.
As I said, if you are reading this as a working Network Engineer, you likely already have this covered. But remember, new people are coming into networking who likely have many of the software-centric skills below but don’t know networking, so use this section as advice on how to bring new people up to speed for basic networking knowledge and networking certs.
1. NGNE Development: 100 Level Subjects
After learning networking basics, some software development tools and practices are going to be incredibly helpful and you should get familiar with them. Increasing literacy in software development processes and models by more-and-more people means that the processes and tooling knowledge (and the tooling itself) is becoming more common in NetOps. This is where the puck is going. The trick is to figure out how much of it you need now versus a year from now and the year after that, and so on.
Here are areas where you need some minimum familiarity in the new world of networking. Again, in future posts, we will give much more detailed descriptions of what you need to learn in each of these sections, and where to find it, but for now, here are the 100 Level Subjects for any NGNE to start with. If you have already learned any of these, you have an excellent headstart, but nothing in this section can be skipped.
Linux - the Industry Operating System
Linux is the OG computer operating system that lies behind many production applications and systems today. Learning Linux will get you familiar with operating system (“OS”) basics and concepts such as navigating file systems and viewing,creating, and editing files. Linux is also the underlying OS in many Network Operating Systems (“NOS”) such as Junos, SRLinux, and others. It is also the underlying OS used by MacOS. Cap this Linux learning with some shell scripting and other constructs in Linux (e.g., bash scripts, crontab) will also introduce you to the basics of automation, which definitely comes up soon.
Using GitHub?
GitHub is a source code control system (SCCS) that allows teams of people creating software, documents, scripts, etc. to contribute to and edit a codebase, as well as check each others’ code. You could substitute GitHub for another SCCS (e.g., Gitlab or others) if you know your environment is going to use that alternative SCCS, but GitHub is widely used in software development projects and increasingly used in network operations.
Python Scripting?
Python scripting is by far the most popular language today used for writing scripts to automate network operations tasks. Let’s state clearly that the goal here is not to make you a full-fledged Python programmer. Rather, you should start here to understand the basics of automation for NetOps. That being said, if the programming aspects of Python really intrigue you, Python is used in many other technical disciplines (e.g., Data Science), and this can broaden your skill set and expose you to other areas for future career growth.?
You could substitute Python for another language suited for automation (e.g., Go is rising in popularity for Network Automation), but Python is a practical choice due to its popularity and wide ecosystem of support.
Using a Cloud Platform?
Cloud services are here to stay, and knowing how to use them and understanding networking between physical on-premises devices and the cloud is an absolutely necessary skill for you as a NGNE.?
AWS is very popular and a good place to start, but there are others to choose from, such as Microsoft Azure, Google Cloud Platform (GCP), Oracle Cloud Infrastructure (OCI), and others. It’s probably best to pick the one that your organization is already committed to, or, better yet, the one they are moving to soon.
Jumping into the cloud exposes you to a whole suite of services for IT. For NGNE, we suggest that you become familiar with the networking constructs as a place to start. You might be tempted to think that cloud networking is just like “trad” networking, but this is definitely not the case – cloud networking is highly abstracted away from hardware, and it’s very different in many other ways. It will be important to understand the differences. You can certainly go much deeper on cloud services than just networking, but we suggest at a minimum that you know the definition and utility of the basic functions in the cloud such as compute, storage, firewall, and database.
Virtual Lab Environments
There are a growing number of tools (sometimes free or low cost) that let you set up virtual environments for networks. These are critical for you as an individual to learn and test specifics of different types of network elements. Some examples of these are Containerlab, EVE-NG, and Cisco Modeling Labs (CML).
AI-Enhanced NetOps Skills
AI is definitely a class of tools you need to understand and understand how to use. There is clearly a great deal of hype around AI, and including it here is not an attempt to jump on the bandwagon. In fact we want to help you burn through the hype haze and understand what AI tools and use cases provide real utility in NetOps. Our real goal in this section is show you how to use AI as a key differentiator to springboard your career.
AI can both help you learn about many of these NGNE topics and allow you to automate many NetOps tasks and operations. You can start playing with your favorite chat-based AI tools today (ChatGPT, Grok, Gemini, and many others) to help you learn about these NGNE topics.?
You’ll also see many claims of AI enhancing many of the tools you use in NetOps. We’ll be writing more on this topic to show you practical tips on how to use AI techniques for all of the NGNE topics (from basic and advanced), and those posts will help you evaluate the usefulness of AI-enabled NetOps tooling.
A Note on Vendor-Specific Skills
Many of the items above and other advanced NGNE skills could be influenced or driven by specific products you insert in your NetOps stack. You need to add those tools and products to your learning plans accordingly.
2. NGNE Development: 200 Level Subjects
Beyond 100-level topics, there are many more next-level tools you’ll need to learn. Below we give a very cursory preview of these topics, and we'll? write in more detail on them in future posts. In the meantime, here is an overview of the 200 Level topics.
DevOps Principles
The NGNE needs to know how DevOps came to be in the software development and deployment arena, its power and utility as a movement, and where and how its principles can be applied to NetOps to great effect. Some refer to this as NetDevOps and that’s fine, but there are pieces of DevOps that need to be modified for NetOps.
Automation Concepts
I’m cheating here a bit in that I did mention Network Automation above with Python, but here, NGNEs need to go deeper on more than how to automate tasks. You need to develop a way of thinking about why and when tasks should be automated, and how (most importantly) how automation can lead to improved processes versus simply automating a series of tasks in exactly the way you do those tasks today.
Source of Truth (SoT) for Automation and NetOps
You need a baseline of how your network is intended to be configured, and a SoT gives you a baseline for that intended config. It’s a key tool and framework that enables manageability, resilience, disaster recovery, and it is core to Network Automation. There are different ways to approach SoT, and there is active and healthy debate around the nature and function of SoT that you need to be aware of as an NGNE.
Infrastructure as Code (IAC) concepts
IAC borrows heavily from the ways you establish cloud services rapidly via automation and at scale. IAC and tools like Terraform and OpenTofu can help you implement cloud-based network functions rapidly, and can help manage trad network infrastructure in conjunction with tools like GitHub.
Application Program Interfaces (APIs)
This topic may be the most thrown-around term in networking today, but few understand APIs beyond the conceptual and how to use them in the most basic ways. As a network engineer, you understand a device’s CLI, which is essentially a human interface into device config. As we increase the use of automation and employ software processes (applications) to help us configure, operate, and even protect the network, we need to know intimately how to give that software the right APIs into network devices as well as other software. (See the above example of telephony operators being replaced by code)
Ansible
One of the OG network automation tools, Ansible is an open-source software tool that allows you to automate repetitive network configuration tasks. It partly competes with Python and other scripting languages, but remains a viable alternative in many use cases. You need to know the differences between how Ansible and Python do things so you can make informed decisions for environments where you work.
Continuous Integration / Continuous Deployment (CI/CD) Pipelines
In the realm of DevOps and software development, a CI/CD pipeline is an automated process that helps software development teams streamline the writing, integrating, testing, deployment, and more testing of applications. This is tightly-coupled with having separate coding environments for development (dev), testing (test), and production (prod). CI/CD is also useful in cloud development environments, where you can easily clone/copy dev to test and test to prod. CI/CD in Network Engineering can enable you to manage device configs and enable the concept of having a dev, a test network. It is especially useful in automating tests for proposed config and network changes. However, when it comes to trad networking, prod is a network of physical devices, not just a software environment, so you should be aware of how to adapt CI/CD to physical network environments.
Orchestration Concepts
Automating a single task is somewhat atomic, such as automating a single command and capturing that command’s output or upgrading the NOS on a device. Many NetOps activities require the chained execution of atomic automations with coordination of automations based on outputs of multiple steps. You might also want to orchestrate automations across multiple IT domains, such as automated coordination of activities between applications and the network, or an IT service portal that may need to create services with dependencies between tech domains. Orchestration can help you automate the automations–reliably and at scale–within just the networking domain as well as across multiple IT domains.
Workflow
With the potential for automations and orchestration of those automations to get increasingly complex, we need to think more explicitly about the processes we want to automate (sometimes without even thinking about them) and how we implement them for our NetOps tasks and operations. Thinking deliberately about workflow, with the mindset of being open to reengineering the manual processes you are automating, can lead to important NetOps efficiencies.
3. NGNE Development: 300 level
Similarly, there are next-level advanced topics that NGNEs will need to know about. Again, we’ll write more about these later, but as an introduction, these topics include:
Multicloud Networking (MCN)
You’ll eventually need to (or may already) use more than one cloud platform, and provide connectivity from your on-premises infrastructure to those cloud platforms, and probably even between those clouds. In networking, we’re pretty fortunate to have standards-based protocols that govern how network devices from different vendors and different open source projects interact with each other. The rise of the disparate cloud platforms has not been governed by the same types of standards for interoperability. Consequently, AWS has one way to instantiate and manage its services, Azure has others, and so on. This makes networking between different clouds (and trad network gear) a challenge; MCN is a set of concepts and tools that are helping with this.
Observability and Telemetry
For a long time, we simply had SNMP and syslog messages to figure out how network elements and traffic were behaving (or misbehaving). These were important starting points, but as we became more and more dependent on Internet-based applications for critical services (starting with voice and telephony), we needed better visibility and observability of the real-time status of network infrastructure, traffic load and mix, and the resulting performance and behavior. The growing field of observability and telemetry tools, and the open source tools that let you put useful graphic views of that info together, enable you need to know what is happening with your traffic, see how it changes over time, and how certain network conditions should guide, trigger, and leverage automation and orchestration.
Digital Twins
Network modeling has seen limited use over the past ~30 years, largely due to dependency on vendor-specific hardware. Today, being able to separate the control plane software from the hardware-focused forwarding plane, along with the increasingly-inexpensive access to compute resources, have combined to enable us to emulate network control planes at larger scale. Combining this with time-series database technology and observability tools, we’ve seen the development of digital twin technology that can allow us to see the state of a network over time, and use observability data to predict future state of the network. You may have little exposure to these technologies and concepts, but we predict they will become very useful in the very near future.
Security
It’s not fair to just say “Security”. This is an incredibly broad and complex set of technologies, and it is constantly evolving due to the creative nature of bad actors continually trying to do new bad things in new and clever ways. Suffice it to say that NetOps and SecOps already have significant overlap and that overlap is always increasing. It will be a great operational benefit to view them together and think holistically about how NetOps and SecOps interact and overlap.
Other NetOps Tech Stack Adjacencies
Security isn’t the only adjacent tech that can impact NetOps. Other tools and silos will also need to be considered. Some of these are optical networking, the collaboration tools your team uses, your ticketing software, procurement, and many other parts of your ops tech stack are highly interdependent. Learning to view them as a system will lead to other efficiencies.
Continuous Evaluation
We live in a time of very interesting accelerated innovation, where AI is only one of the accelerators (albeit one with significant promise). This means new tools (both open source and commercially available) are becoming available at a rapid pace. You need to adopt a mindset of regular, if not continuous evaluation. This means intentionally and proactively keeping eyes out for what tools NetOps and other adjacent parts of your Ops stack are becoming available. You will need to build processes and criteria for evaluating those tools. The “criteria” piece of this is important – just because something is new doesn’t mean it will be helpful in your environment.
4. NGNE “Electives”: 400 Level Subjects
Some new things are already solidly entrenched in Networking Operations at organizations that are? and not going away. While not all of these should be a required topic, here are items on our radar that are good to know and emerging items that can help you in NetOps. For many of them, if they are already on your radar, are very interesting to you, or are already in use in your organization, you should add them to your curriculum.
Containers, Docker, and Virtual Machines
Containers have been around in Linux for a long time. As with other software constructs, software processes delivered via containers created and managed with Docker are key tools to understand. Many of the software and automation tools noted above are delivered in containers, so you will run into containers somewhere along the way and need to know how to use them.
Virtual Machines (VMs) are a predecessor to Containers, though they are considered a more resource-intensive and less graceful way to deliver software. You will likely run into vm-based software and should be aware of VM concepts and also know how VM and Containers operate differently.
Kubernetes
Kubernetes (k8s) is a tool for managing and automating the deployment, scaling, and operation of containers. While k8s also comes primarily from the software deployment world, there is a growing movement to leverage k8s constructs for network automation. If you have the time and interest, it’s worth learning some basics about k8s and see how it could be used for Network Automation and how networking is implemented in k8s.
Understanding Open Source?
More and more can and is being done with open source software for NetOps. Some understanding of how to use, contribute to, and get support for open source software is a useful area for study.
Open Networking
Open Networking, also referred to as “disaggregated networking”, where the NOS is purchased or obtained separately from the hardware, is also an increasingly-viable option for networks. Some NOSs are open source, such as SONiC, and others are vendor-developed and supported. They tend to run on “merchant silicon” platforms, often based on network ASICs from Broadcom, but other manufacturers are gaining momentum (e.g., Marvell). Integrating and operating network elements in this mode requires a level of knowledge on how they are different from “shrink-wrapped” products, as well as more responsibility for making the hardware and software work together. A well-rounded network engineer should know at least a little about this area.
Web Assembly
The computing and software industry evolved from monolithic programs, to virtualization via VMs, and to Containers – Web Assembly has potential for being the next step in delivery and operation of programs across disparate infrastructure.
Next Steps to Becoming a Next-Gen Networking Engineer
As mentioned above, you’ll need to tailor all these topics to your environment. You’ll consider vendors in place and on the horizon, as well? priorities in your NetOps team and what interests you. Here are some general approaches to using the NGNE learning roadmap:
Go Deep on Topics in Order
You could use the above learning roadmap laid out as-is. We think this works well with the topics listed in NGNE Levels 100-300. You could even cover a smaller number of subjects at a time.
Go Wider: Overviews of More Topics at a Time
You could do more of an overview of the topics and come back to each one for more depth as needed or as interest dictates.
Triage: Do What You Need To Do
You could dive into the things you need to know first in any of the levels, driven by your job, or by your curiosity, regardless of where they sit in the NGNE levels. Do what you need to do.
Study Groups
To supplement any of the above modes of NGNE learning, getting together with one or more other people with similar interests can be really helpful. There are organizations, podcasts, and individuals that can help specifically with learning in groups. You can also just reach out to people you know who are in the same situation.
Start Somewhere
Whatever you do, you need to work through the topics in a way that balances your work needs, the time you have to study, and what interests you the most.
The most important idea to get started is to set some priorities and start somewhere. A year from now, it’s almost guaranteed that something in this list WILL change. Remember, we live in a world of tech disruption. Kubernetes may take over the world, WebAssembly could gain huge momentum, or other topics that aren’t even on our radar yet will emerge.
TNOps 2025: More To Come!
We have some big things in the works from TNOps for 2025.?
NGNE Content
In addition to fleshing out the NGNE levels and topics above, we’re working through ways to point you to great resources on the above topics, whether you want to pursue self-study materials or you prefer a more guided approach.?
Let us know if you have any stand-out learning and training resources for any of these NGNE topics. Conversely, wherever you see gaps of excellent training materials, let us know. We are already working on content to fill gaps we see, but we want to hear from you. Leave comments or message us letting us know where you’d like to see more work done in training material.?
Org Change: NetOps Needs More from IT And Tech Ops Organizations?
As noted at the beginning of this piece (wayyyyyy at the top), we want to call out again an important issue here. The need to upskill and adapt is often put squarely on the shoulders of the Network Engineer, and you are often expected to use your nights and weekends to skill up.
If that’s the organization you work in, you may just need to just suck it up and learn on your own time, but we don’t think that’s the most effective or fair approach.?
NetOps and IT organizations in general benefit greatly when network engineers gain new skills and bring those capabilities to the job. The best organizations realize this and work proactively to upskill and cross-train their staff.
We’re working with organizations to apply TNOps principles to NetOps org improvement and development, to support network engineer tech and career development, and to also focus on improved NetOps processes, managing thru change, and helping people who are disrupted by tech changes.
If your organization is ready to start the journey to upskilling your staff or adopting TNOps principles in general, get in touch with us today to talk about how we can help.
In the meantime, stay tuned for more in 2025! You can sign up here to be notified when the next posts in this series are available where we’ll go into more details in 200 and 300 level subjects, and start to put together the best list ITW of where to get the training and upskilling you need to master the revolution that is happening in Network Operations!