Skullgirls Back End & LiveOps
The Engagement
Skullgirls Mobile is a popular fighting game franchise with millions of players worldwide. The mobile version requires a robust and scalable backend infrastructure to handle massive amounts of player data, real-time transactions, and game state management.
As the game continued to grow, the backend infrastructure needed modernization to ensure optimal performance, cost efficiency, and future scalability.
The client needed expert DevOps and back end specialists who could take over their complex system of servers, build pipelines, and databases and modernize it without disrupting the player experience.
Autumn Games partnered with Double Coconut to migrate their legacy infrastructure to modern cloud solutions.
As the game continued to grow, the backend infrastructure needed modernization to ensure optimal performance, cost efficiency, and future scalability.
The client needed expert DevOps and back end specialists who could take over their complex system of servers, build pipelines, and databases and modernize it without disrupting the player experience.
Autumn Games partnered with Double Coconut to migrate their legacy infrastructure to modern cloud solutions.
Timing
3-4 months for complete migration - ongoing infrastructure management and liveops feature development
Team Composition
- DevOps Specialist
- Server and Database Architecture Expert / Tech Director
- Senior Web Angular Engineer
- Backend Systems Engineer
Our Contribution to This Game
The Problem
We inherited this thirteen-year-old game and had a few fun problems to solve:
- Legacy database infrastructure: The existing MongoDB sharded cluster was running on outdated version 3.2, creating security vulnerabilities and limiting access to modern features and performance improvements.
- Massive data migration complexity: 8TB of player data needed to be migrated without data loss or extended downtime that would impact millions of active players.
- Zero tolerance for disruption: Any extended downtime or data loss would directly affect player experience, potentially leading to player churn and revenue loss.
- Infrastructure cost optimization: The existing infrastructure was not optimized for cost efficiency, and the client needed a solution that would reduce operational expenses while improving performance.
- Finally and most important: We had to help create APIs and systems to suport new high-scale features such as guilds, deeper analytics, and a new battle pass.
- Unstructured and incomplete admin panel / CMS.
Our Solution
Our team took full ownership of the backend infrastructure and executed a comprehensive modernization strategy:
- 1. Zero-Downtime Database Migration We engineered a sophisticated migration strategy to upgrade the MongoDB cluster from version 3.2 to version 6.0 while handling 8TB of live player data. The migration was executed in carefully planned phases with only two brief maintenance windows: one 10-minute window and one 30-minute window. Throughout the entire process, zero data loss was achieved, preserving all player progress, purchases, and game states.
- 2. Cloud Synchronization and Validation We implemented a dual-cluster synchronization strategy using MongoDB Atlas. Over several days, we achieved 99.999% data synchronization between the legacy cluster and the new Atlas cluster. This sync was maintained for an extended validation period to ensure data integrity and system stability before any production cutover.
- 3. Performance Maintenance During Migration Throughout the entire migration process, the legacy cluster continued operating at full capacity with no performance degradation. Players experienced no lag, downtime, or service interruptions during the transition period, demonstrating our expertise in managing live production systems.
- 4. Infrastructure Testing and Modernization Per the client’s requirements, we orchestrated a comprehensive test run of the new infrastructure. We bootstrapped a test ECS cluster configured to communicate with MongoDB Atlas, validating our automated scripts for NAT gateway provisioning and VPC configuration. This testing phase ensured seamless communication between game servers and the cloud database infrastructure before production deployment. list item
We helped build a modern and cohesive web-based dashboard.
- Advanced game design, merchandising, and other live-ops features
- Atomic roles tuning ability
- Importing/exporting/transferring data between different environments
- Handles mobile builds and manages maintenance / releases
- The Angular-based dashboard runs on AWS lambda, so if not being used it doesn't have any cost.
The Results:
Our comprehensive backend modernization delivered exceptional outcomes:
• 30% performance improvement: Players experienced faster load times, reduced latency, and smoother gameplay through optimized database queries and modern infrastructure.
• 30% cost reduction: By rightsizing infrastructure, leveraging cloud-native solutions, and optimizing resource allocation, we reduced the client’s monthly operational costs by nearly one-third.
• Future-ready architecture: The modernized infrastructure is now scalable, maintainable, and positioned to support future growth and feature development.
• Seamless transition: Despite the complexity of managing 8TB of data and millions of active players, the migration was completed with minimal disruption—just 40 minutes total downtime across two maintenance windows.
Our comprehensive backend modernization delivered exceptional outcomes:
• 30% performance improvement: Players experienced faster load times, reduced latency, and smoother gameplay through optimized database queries and modern infrastructure.
• 30% cost reduction: By rightsizing infrastructure, leveraging cloud-native solutions, and optimizing resource allocation, we reduced the client’s monthly operational costs by nearly one-third.
• Future-ready architecture: The modernized infrastructure is now scalable, maintainable, and positioned to support future growth and feature development.
• Seamless transition: Despite the complexity of managing 8TB of data and millions of active players, the migration was completed with minimal disruption—just 40 minutes total downtime across two maintenance windows.
What Success Looks Like:
Not only did we successfully take over and operate a complex production backend system without any issues, but we transformed it into a modern, cost-effective, and high-performance infrastructure that positions Skullgirls Mobile for continued success by having a control center to run liveops.
Not only did we successfully take over and operate a complex production backend system without any issues, but we transformed it into a modern, cost-effective, and high-performance infrastructure that positions Skullgirls Mobile for continued success by having a control center to run liveops.