Mountkirk Games Case Study: Scaling to Millions on the PCA Exam
If EHR Healthcare is the "legacy migration" challenge, Mountkirk Games is the "born-in-the-cloud" global scale challenge. In this walkthrough, we'll break down how to design a backend that handles millions of concurrent players without breaking a sweat—or your budget.
The Mission: Global Gaming at Scale
Mountkirk Games makes mobile games. Their latest hit is exploding globally, and their current backend (running on another cloud) is struggling with the massive, unpredictable load. They need a solution that is highly available, globally scalable, and provides real-time analytics.
Key takeaway for the exam: Mountkirk is about managed services. You shouldn't be suggesting Compute Engine VMs for this if a serverless or containerized option exists.
The Database Dilemma: Spanner or Bigtable?
This is where most students get tripped up. Mountkirk has two distinct needs:
1. User Profiles & Scores (Cloud Spanner)
For global consistency and relational data (like user accounts and high scores), Cloud Spanner is the hero. It gives you the ACID compliance of a relational DB with the horizontal scale of NoSQL.
2. Time-Series Metrics & Game State (Cloud Bigtable)
For high-throughput, low-latency telemetry data coming from millions of mobile devices, Cloud Bigtable is the go-to. It's designed for massive write volumes—exactly what you get when millions of users are playing at once. Check our database flowchart for a full comparison.
Exam Hack: If the question asks for SQL and global scale, pick Spanner. If it asks for huge write throughput and NoSQL, pick Bigtable.
Real-Time Analytics Pipeline
Mountkirk wants to see what's happening *right now*. How do you build that pipeline on GCP?
- Ingestion: Cloud Pub/Sub handles the incoming stream of events from millions of clients.
- Processing: Cloud Dataflow (streaming mode) cleans and transforms the data on the fly.
- Storage: BigQuery is used for the final analysis and long-term storage of business intelligence.
Zero-Downtime Infrastructure
Because games never sleep, Mountkirk needs GKE clusters spread across multiple regions. You'll want to use the Global External HTTP(S) Load Balancer to route users to the nearest healthy cluster, ensuring the lowest possible latency for a smooth gaming experience. For an in-depth comparison of container options, read our GKE guide.
Frequently Asked Questions (FAQ)
Why not use Cloud SQL for global high scores?
Cloud SQL is limited to a single region for writes. For a global gaming scenario like Mountkirk, Cloud Spanner is required to provide consistent, globally available writes across multiple continents.
What is the benefit of using Pub/Sub in this architecture?
Pub/Sub acts as a buffer between millions of game clients and the processing backend. It ensures that the system can handle huge bursts of traffic without losing data or crashing the analytics pipeline.
How does Mountkirk handle analytics cost-effectively?
By using BigQuery for long-term storage and complex analysis. BigQuery follows a pay-per-query model and allows for cost-efficient storage of massive amounts of historic game data — see our cost strategies.
Can they use serverless compute?
Absolutely. For small, event-driven tasks like user registration or processing high scores, Cloud Functions or Cloud Run can be used to further minimize management effort.
How to secure global game traffic?
Implement Cloud Armor with the Global Load Balancer to protect against DDoS attacks and SQL injection, ensuring the game stays online for users in every region.
Summary Checklist
If you see Mountkirk Games, look for these keywords in the correct answer:
- Horizontal Scalability (not vertical!)
- Managed Services (GKE, Cloud Functions)
- Multi-regional Deployment
- Managed Instance Groups (MIGs) for compute
Practice makes perfect!
We have specific questions based on the Mountkirk Games scenario in our Mock Exam pool.
Take a Full Mock Exam