Google Professional Cloud Architect Certification (PCA) – Case Study Questions

Google Cloud Architect Certification (PCA) exam is considered to be one of the most valuable cloud computing certifications today. This certification is for professionals seeking to demonstrated that they could design, develop, and manage solutions using Google Cloud. In this exam, some of the questions refer to case studies that describe a fictitious business and a solution concept. These case studies are intended to provide additional context to test skills for designing and implementing solutions. This article lists some of the possible questions for practice that are based on formal case studies :

• Mountkirk Games
• Dress4Win
• TerramEarth
• Mountkirk Games

Mountkirk Games makes online, session-based, multiplayer games for mobile platforms – all hosted on leased physical servers. They had had problems scaling their global audience, application servers, MySQL databases, and analytics tools. For their future game releases, they want to host the game backend on Compute Engine to capture streaming metrics, run intensive analytics, and integrate with a managed NoSQL database.

Refer to the case study link for technical environment list referred in below sections.

Scenarios :

1. MountKirk game statistics are written to files and use ETL tool to generate reports using MySQL database. In the new world, the technical requirement is to process incoming data on the fly directly from the game servers. The possible question in this scenario can be about the solution approach for migration plan utilising BigQuery & Cloud DataFlow.

2. The database used is MySQL, whereas in the future, their technical requirement is to use separate databases for game activity (10TB size) and transactional data (user profiles & game state). Questions can test the capabilities of designing the architecture for handling database workloads. BigTable for time-series data and BigQuery for historical data queries can be the choice considering the technical and business requirements.

3. Both the business and technical requirements specify that dynamically scale up or down based on game activity. Slowness will lose them, players. It is essential how the backend platform deployment architecture can be designed to address this requirement. A deployment solution can be template-based, multi-zone managed instance groups with a Layer 4 load balancer for routing traffic to different servers.

4. One of the requirements for their analytics platform is to process data that arrives late because of slow mobile networks. This requirement needs significant testing to verify the resilience of the mobile network latency. This testing solution should look at possible ways to process the data from mobiles connected to data networks of varying speeds. Here is a good reference for some tips on how to answer this type of questions.

5. Real-time analytics platform for MountKirk’s new game can use a combination of technologies Cloud Storage, Pub/Sub, Dataflow and BigQuery. Questions were testing skills on using Pub/Sub for data ingestion, low latency data processing using DataFlow and data warehouse BigQuery for immediate analysis via SQL or visualisation tools. In addition to this, other scenarios for questions can be application environment deployments, CI/CD pipelines and troubleshooting of specific errors with new infrastructure.

Dress4Win

Dress4Win is a web-based company that helps their users organise and manage their wardrobe using a web app and mobile application, hosted in a colocated data centre. In order to meet its growth in business, the company is committing to a full migration to a public cloud – starting with their development and test environments. They are also looking at building a disaster recovery site because their current infrastructure is at a single location.

Refer to the case study link for technical environment list referred in below sections.

Scenarios :

1. Migrating MySQL database to cloud equivalent will be a key exercise; however, as per the business requirement, they want to minimise downtime and performance impact. Questions can focus on the migration plan for database workloads with the restrictions given in the requirement. As only development and test environments are only migrated first, the solution can address ways to set up a replica slave instance on cloud and asynchronous replication with the master instance.

2. As there is a requirement to keep the solution to address low latency for global customers, questions can be on the deployment of web & transactional data layers (Nginx, Tomcat, MySQL), server machine types, load balancers, approach for production deployments and scaling up/down of the capacity. These questions potentially test the candidate skills in using global load balancers, regional scoping and managed instance groups etc.

3. For the production infrastructure, identify the resources as equivalent to Tomcat, Nginx as application servers; MySQL database; RabbitMQ for messaging; Hadoop/Spark for analytics and NAS Storage. The equivalent GCP resources need to be configured, keeping the business and technical requirements in mind.

4. The requirements specify improving security and also data encryption on the wire, and at rest, questions on solutions for these also can be expected. Building experience in security-specific resources like Google Groups, IAM Roles and encryption methods will help to answer these questions successfully. In addition to this candidate can expect general questions on how to address auditing and compliance requirements.

TerramEarth

TerramEarth manufactures heavy equipment for the mining and agricultural industries. TerramEarth’s vehicle collects data from fields and stores it locally on the vehicle. This data is later offloaded for analysis using gzip, FTP with the help of a Python script and then stored in a PostgreSQL based data warehouse. Data analysis and improvement of parts take 4 weeks; this makes the downtime of vehicles for the same duration. Executive management wants to reduce the unplanned downtimes like this to a week and expand beyond a single data centre.

Refer to the case study link for technical environment list referred in below sections.

Scenarios:

1. The key requirement for the business is to reduce the downtime and wartime for customers for parts. This has been defined as reduction buy three weeks in the requirements. Candidate can expect a question on what to focus on for achieving this level of reduction. There are immediate areas that can be looked at here to improve the service – like file transfer methods, reliability of file transfer, ETL, metrics reporting and cellular connectivity.

2. The overall solution should address ways to retrieve data from the vehicle in a standard way that can be used by dealers and partners so that the inventory of parts can be quickly organised rather than waiting for the vehicle to be available for service. This is an area where the candidate can expect questions on the Google Cloud EndPoint APIs and its OAuth compatible authentication methods for dealers and partners.

3. In addition to the questions related to the cloud infrastructure, the examination may cover questions related to collecting telemetry data from servers and servers of connected vehicles and use it for training models using machine learning. These data can be compressed within vehicle’s computers as hourly snapshots and upload to store it into GCS Coldline buckets. This approach will reduce the cost of storage.

Summary

PCA certification is for cloud architects who want to build sophisticated cloud solutions and services using GCP. PCAs must develop a thorough understanding of cloud architecture in general, and the Google Cloud Platform in particular. Those who hold this credential can design, develop and manage dynamic Google Cloud Platform solutions to meet business objectives that are robust, secure, scalable and highly available.