What is your return of investment rate when learning network topics?

Sunday morning meditation, take it with a grain of salt.I am reading about Graph DBs and the below came to my mind.

Hi guys

My question stems from the fact that I am now learning network programming and I realize that my learning is guided by concrete needs and problems that I can solve and I learn knowledge that I can apply without facing an uphill battle (I am still an one man team :-) when it comes to programming. I do not need to convince anybody about what I want to do, I take the decisions, work on the problems and present the results. That does not really go the same way when it comes to other things I have learned about networks since I work in a large environments with multiple teams

​

Over the years I learned a lot of things about networking. Looking back I can say I am under the Paretto's rule (20/80) and I had the opportunity to apply little of the deep knowledge that I acquired in certain areas. Luckily the knowledge is preserved as notes which I revisit as needed but it fades because of not being used. That happens because either you don't need it or the environment you work in is bureaucratic or it tries to keep things simpler (you need to design things that have to be easy to operate and maintain, often with impact of performance and other important factors) .

I must say that learning network programming and related topics (ex: working with databases, frameworks, APIs, libraries, GIT etc) give me more satisfactions than learning the network topics themselves. Paradoxically, this new learning benefits from my past learning because I can now make use of some of the knowledge I acquired in the past in areas like sysadmin (linux) . These days if it would not be for network programming I would probably never go back to use some of the knowledge I have about routing protocols, their routing databases, the protocol machine state and other things you rarely get to use as engineer or even in an operational role. The new SD-whatever paradigm, while fascinating, takes us even farther from these things, hiding complexity and details that the new breed of network engineers don't apparently need anymore (that is my perception) The large mass of roles available going forward won't need these details anymore. Looking back you can see how the technologies evolved and built on top of each other from the beginnings of the Internet and TCP/IP protocol units to hiding the network behind an AWS, Azure or some managed SD-XXX solution ...

So what is/was your ROI from learning topics like the ones in CCIE R&S (one of the few curriculum that forces you to be well rooted in the fundamentals) . Please don't count "it opened many doors for me" because that will mislead the new guys. It opened the doors for you because of the market perception of the cert not because someone did have the skills and knowledge to measure yours :-) . The job requirements now cover Network programming/more like scripting, Linux knowledge (every damn network device has a Linux kernel these days) Virtualization and others which force us to get into to spaces that were never ours before. To understand better where I am driving with these think about the serial connections and their protocols and how critical they used to be a good while ago and how forgotten and little used they are today. I merely had a glimpse at the end of that era, I started with mostly Ethernet but .... you get the idea, I have the vague feeling that deep knowledge about these things will not be needed for many of the network engineers of the future and because of emerging technologies they will be forced by the market to learn more about the abstraction on top of these things. If you need an analogy we stopped replacing transistors and circuits, the art of using an oscilloscope to trace signals, transitions and others is the attribute of very few, it is cheaper these days to replace the module (or it will be cheaper to deploy things in AWS than to maintain your own network) Among other things I am learning ACI these days. I know it was a disappointment for many of you but I believe it will grow to be a very solid product. It will have competition (I hope) but looking at the change in paradigm you can't stop of thinking about the above

​

Just to be clear, I am not panicked or frustrated or anything, I accidentally morphed a while ago without even realizing I started using Python for something else and later I started moving with the flow because I could.

I am just thinking out loud and I wanted to see what others think.

​