It’s been 15 years since I completed the undergraduate course. Neither during the course nor after it did I read a technical book from cover to cover. Then, a few years ago, in October 2018, I came across OSTEP (Operating Systems: Three Easy Pieces) and, in an impulse, purchased its physical copy. It stayed in my home library as “to read” book. I remember reading a few pages when the book was first delivered. But nothing beyond that. However, ever since I read its few initial pages, I was aware that the authors’ writing style matched my learning style in that:

  • It wasn’t too formal
  • It felt like the author was having a conversation with the reader rather than trying to preach/sermon

Moreover, I wanted to pick it up to understand OS better because, as a teen, Linux got me very excited, but I could never pursue it as much as I should have

Picking it up, at last Link to heading

Fast forward to November 2024 and I came across this tweet:

It felt like a signal from the universe to read the book because, back in November 2024, I had addressed a few career and personal tasks, and in my head I was clear about what I wanted to do: become an engineer I could myself respect. To be able to be such an engineer, I had to learn, re-learn and develop an understanding of few CS subjects that I had either forgotten or never studied in the first place. E.g., I had forgotten most of what I had learned in Operating Systems, Data Structures, Database, etc. classes, and never studied Algorithms in the first place!

Now, in May 2025, I have finally finished reading the book while juggling family and work responsibilities. I never thought I’d finish the book! It’s little over a month late than when the book club finished reading it, but that’s not a problem. Finishing the task has surprised and satisfied me the most!

What did I learn? Link to heading

To be honest, I wasn’t confident about finishing this because I have a history of starting projects/tasks that don’t cross the finish line. Especially the technical books I have started always remained incomplete so, I was only annotating the book here and there.

It would be beyond dumb of me to say that I now know everything in that book perfectly. My goal was to understand topics in the moment and not care/bother/stress about remembering things any longer. And, indeed, almost each time I left a chapter midway and came back to it after a few days (even 2 days), I had to surely re-read it.

I once wrote this post titled “Stack, Heap, Go” as a “conversation” with some colleagues. However, the truth is that it was a question I was asked in an internal interview at Red Hat when I applied for the KubeVirt team. It was the very first question I was asked over the multiple rounds of interviews I gave before eventually being able to clear them and joining the team. But when I fumbled in that question, I was 0% hopeful of being able to clear the interviews. It further cemented the impostor syndrome thought process in my head.

The takeaway for me is that an OS does A LOT of stuff under the hood that we don’t give much thought to. But we need to understand those to write better programs. Well, I always knew it does a lot under the hood, but the specifics became clearer as I read OSTEP. Some of these include:

  • scheduling a program on a multi-CPU system
    • so many approaches, so many intrecacies
      • well this is kind of true for everything that OS does
  • managing the memory of programs
    • CPU cache and multi-CPU architecture makes this an even more challenging problem
  • the stack and heap that I couldn’t talk about in the interview were all over the place in the section on “memory virtualization”
  • concurrency is crazy with C
    • initially I was excited about this section and dismissed the author’s note in the “dialogue” chapter about this being a topic causing headaches. Reason for this was having written some concurrent code in Go.
    • I quickly realized why Go’s concurrency model and syntax are so very loved by polyglot programmers.
    • and I indeed ended the section with a headache.
  • although the section on concurrency felt challenging, I’m positively excited about learning further.
  • the section on storage and file system felt a lot of fun!
    • probably because it was the last section so my mind remembers it the most! Remember I took almost six months to finish the book.
    • it might also be because authors are doing their own research around this topic, so their explanations here might have been more fun than they were in other sections.
  • the amount of things happening to make an I/O request happen blew my mind away!
  • filesystems are so easily taken for granted, but it’s a whole field of research on its own. I also stumbled upon a podcast episode about ReiserFS while reading this section which added some context.
  • the part about data staying in page cache before being actually flushed/written to the disk is also really interesting! I first learned about it in this talk about questioning the design patterns of storage engines. I was at the conference so I saw it live and I had started reading OSTEP only a few weeks ago, so all the while I was thinking that this topic about fsync() system call is going to come up in OSTEP. The under the hood working of I/O is covered pretty amazingyly in the book!
  • the section on distributed systems is, thankfully, short.
    • I did the CKA and CKAD certifications of Kubernetes in November 2020. When I shared it with my team at work, a colleague recommended me reading the Designing Data-Intensive Applications or DDIA. I didn’t say it to him, but, after looking at its table of contents, I felt too imbecile to read it. Now, I know I have a huge bout of impostor syndrome, but this book really felt way above my level. After being able to addres a few issues in 2024, I was clear that I want to first learn things from a single system’s point of view, i.e., Operating Systems and Data Structure & Algorithms before thinking about Distributed Systems. This is because, for example, an OS maintains state of various processes and makes sure that each of them gets its fair share of time to run on the CPU. OS does so much at a single system level that, I think, can be translated into what a Distributed System does over connected machines. I want to learn these better before learning Distributed Systems.
    • yet the short discussion of distributed systems is enough and fun to spark the interest to learn more. Now I know that I’ll read DDIA at some point in time.
  • best and most accurate answer to a question about how to do something — anythnig — is “it depends”. The computer is full of binary - 0 & 1. But reality is always between 0 & 1. It’s like CAP theorem which states that one can have only two of consistency, availability, and partition tolerance. At a real job this often translates to being able to have, at max, only two of good pay, great WLB, and interesting work. 😉

What’s next? Link to heading

Looking at some of the commentary from folks who have read OSTEP as well as what authors themselves mention a few times in the book, this book aims to give only an introduction of a lot of topics. So it is up to the reader to dive further into what piques their interest. For me it would be scheduling algorithms, concurrency, and file systems.

But at the moment I want to focus on Data Structures and Algorithms because:

  • it forms almost the absolute basics of the Computer Science domain
  • companies often grind candidates on this topic

I had the privilege of discussing this with my mentor at work and he explained how above two are different in nature. One could learn the Algorithms like they did in university by covering many algorithms, yet it won’t help secure job at a company that grinds on that topic because those interviews are focused on a few algorithms and questions go into some depth of these. At the moment, I’m clear about doing former and have picked up the book Algorithm Design Manual by Steven Skiena. I found it via Teach Yourself CS where it’s mentioned as a book that:

typically succeeds in fostering similar enthusiasm among his students and readers

I’m more interested in enjoying learning it right now just like I enjoyed OSTEP and Let Us C which made me enjoy learning C language at one point. The book we used in university to learn C made it, to say it mildly, a torture to learn the language!

That’s it! Link to heading

Hope to write a post about finishing the book on Algorithms. :)