Let me share my experience of obtaining Google PCA certification in May 2024.
I have already invested a lot of time (and some money) into learning AWS cloud platform. The prospect of learning yet another platform was not appealing to me, to say the least. My last two clients were heavy users of the GCP platform, though. You can go a long distance on generic cloud/IT knowledge and pieces of information you pick up on the job. Learning on the job is the best, but the longest way to learn. At some point I would need to take a holistic, academic training, I thought.
By a lucky chance, my job offered me time-limited “Get certified” program. It included a free exam, 90 days of Quicklabs/Cloudskillsboost subscription, several instructor-led trainings and self-paced trainings. Greediness beats laziness, so I took the opportunity and started preparation.
I don’t think the program offers exclusive content – you can get the same training, the same labs in the public space. Quicklabs/Cloudskillsboost subscription is not expensive. What the program does though – it forces commitment. You have a certain number of weeks to complete a section, you have a certain number of days to book the exam, and you have a certain deadline to take it. In the modern world, there is no lack of information. There is a lack of time and commitment, so I’d say sense of urgency and deadline were the main values of the program for me.
You can get the same benefit by courageously booking the exam in 3 months’ time. Do not research much, do not poke around. Book the exam first, then start preparation. Having a deadline is a strong motivator.
You have a much bigger risk of abandoning certification, than failing the exam.
From the beginning of the journey, I was worried that my AWS knowledge would interfere with GCP knowledge. I thought I’d start conflating products and concepts and end up with worse knowledge of both. To avoid that I have made it a point to clearly compare AWS and GCP in my training notes. As soon as I meet a new GCP concept or product – I would recall and note the similar concept or product from AWS. By doing so, I would strengthen my knowledge of both, as contrast and comparison are great learning tools.
I’ve also used the official reference: https://cloud.google.com/docs/get-started/aws-azure-gcp-service-comparison
Thankfully, there are lots, lots of similarities. It is easier to describe AWS and GCP differences, than similarities. I may do it in another article later. For now, I would highlight that the main difference is in the IAM approach. For GCP the unit of logical segregation is Project. AWS offers Accounts for logical segregation, which are much bigger units than GCP Projects. You define your GCP access policies in a hierarchy, and the lowest node of the hierarchy is the Project. Permissions defined on the Project level will apply to all resources in that Project. In AWS there is no IAM hierarchy, and you define permissions for resources directly. I am still getting used to GCP Projects, AWS approach feels more natural to me. It is the opposite, perhaps, for people who started on GCP.
The exam scope is broad, high-level knowledge of GCP products and their best applications. It does not go into deep details like nuances of PSA vs PSC, or VPC Service Controls Perimeter design. Typical deployment models like hybrid architecture, are only briefly touched. Nevertheless, it requires a considerable time investment to go through all services, understand and memorise various products and concepts. Conceptual understanding is the skill you will need the most to pass the exam.
The bulk of my training was Cloudskillsboost self-paced courses and Google product documentation. Quicklabs are excellent. I highly recommend them. One of my favourites was cloud deployments with Terraform. As a Security Architect, I don’t need to go down to small Terraform coding details, but being able to read and understand Terraform code helps me a lot. The lab forces you into the deep end of the pool, here is the task – start coding on the spot.
I was reading documentation with “examiner’s eye” – for every piece of content, I would try to predict what the examiner’s question would be. You can guess a lot of questions in that way. For example, it is easy to predict questions about the difference between types of load balancers, or difference between Cloud Interconnect and Peering, or discount schemes.
The labs require good hands-on skills. Thankfully I have a strong sysadmin/engineering background, so dealing with Linux CLI and bash scripting is easy for me. While doing the labs I kept having flashbacks of my early career. Hands-on reinforces high-level concepts, though. Please don’t worry if you don’t have enough hands-on skills – the exam is not CLI focused – just the labs.
In terms of resources, most of my training was from the official sources:
Non-official resources were:
- Whizlabs PCA tests – these are far, far from the real questions, but they help you find gaps in your knowledge
- Pluarlsight GCP architecture and GCP security courses
- Number of Medium articles – nothing remarkable, though
There is a lot of emphasis on PCA Case Studies on the Internet. I have read and re-read the studies several times and I still don’t understand why. They are too ambiguous to give you concrete answers and you don’t need to memorise them (they are repeated at the exam). Concrete context with concrete questions will be in the questions, not in the studies. If anything, these studies are like names of the characters, but the actual play will be in the movie. Don’t over-emphasise them. Jump straight into the movie.
The exam itself was relatively easy. Perhaps due to diligent preparation – I spent 3 months 1-2 hours a day to prepare, or maybe it was due to my technical background or work experience. There were no tricky questions, by tricky I mean questions for which you have the feeling they are designed to fail you. The context and questions were easy to understand, and a lot of answers you can pull via pure logic.
With the remote proctoring option, be aware that there are no toilet breaks. Plan accordingly, as the exam is long.
I will conclude with a few tips:
- The exam is relatively easy and it is worth taking.
- Schedule and pay to give yourself a strong incentive.
- Use your personal email for registration
- Official training resources and product documentation are plenty enough to prepare.
- Read training resources and documentation in examiner’s shoes – would you make an exam question out of this content?