My experience comparing the DevOps philosophies of the 3 major clouds on my way to the DevOps certification "treble"

Simultaneously producing a meaningful article and a humble brag for myself? That's something to be proud of right there.

So, last week I got the final of three DevOps engineering certifications offered by the three major cloud companies (AWS, Microsoft Azure and Google Cloud). All three certifications were a challenge, each presented a similar yet different challenge.

All three companies have a lot of similarities as far of DevOps principles go. However, there are differences in their philosophies reflected in the syllabus, topics and the questions in the exam. In this blog, we are going to explore these differences through the description of each cloud and how they approach DevOps.

AWS

The AWS DevOps Engineer exam in many ways is more extensive than the other two in terms of both the number of questions but also the amount of material present in the exam. This is the exam that probably has the greatest emphasis on Infrastructure as a Code and on auto-scaling.

A lot of the DevOps philosophy of AWS can be defined based on their desire to see a smaller workload become a larger workload and their emphasis on scaling and migrating workloads effectively. This is why they have greater emphasis on things like load balancing and auto-scaling.

Azure

Azure DevOps is about two things: Azure DevOps (the service, not the first thing) and repositories. It is a good system I suppose, since Microsoft owns the biggest repository website (GitHub) and has built-in Azure connections to it. Azure DevOps is sort of an enhanced version of that, which includes artifact storage, build server provisioning and much more.

Azure’s approach to DevOps is the most service-based out of all the clouds. It is simpler yet more pervasive. It is highly likely that even if you haven’t used Azure per se, you have probably used some piece of its DevOps stack somewhere in your work, usually that part is GitHub.

Google Cloud

Google Cloud’s approach is quite different from the other two. There is an emphasis on SRE (Site Reliability Engineering) in this certifications course (largely because Google wrote the SRE Workbook ). There is the usual topics of CI/CD and code pipelines, but there is a greater emphasis on monitoring and disaster recovery.

RPO and RTO are two abbreviations you’ll see often, you see them in the others as well but here they are gold. They are so golden that Google has defined the four golden signals around them: latency, traffic, errors and saturation. All these things are monitored using monitoring services. And when the monitoring metric passes a certain metric, disaster mitigation and recovery is initiated.

Conclusion

They’re all very similar, honestly. The differences represent slight differences in engineering ideology. AWS wants to be more freeform, Microsoft wants to be more streamlined while Google aims to be more strict and reliable.

However, they are all DevOps and the fact that these certifications and the wells of knowledge that surround them exist should show you how important they truly are.

I got the treble and now I’m writing a book about using Python in DevOps (due to be released in January 2024) and it will contain a lot more information on topics such as these. Shameless plug.