AWS for Games Blog
emptyvessel builds global game development pipeline with AWS
This blog post was co-authored by Garrett Young, General Manager and COO of emptyvessel and Wei Ning, CTO of emptyvessel.
In the high-stakes world of game development, success hinges on seamless collaboration across time zones and lightning-fast development cycles. For one ambitious studio creating a next-gen cyberpunk shooter, Amazon Web Services (AWS) became the catalyst for transforming their global workflow and accelerating their time to market.
Founded in 2023, emptyvessel is a remote-first game development studio with 18 developers distributed across the globe from Sydney to Slovakia. The studio is developing DEFECT, a cyberpunk objective-based shooter for PC and console. They chose AWS to improve internal collaboration and increase development velocity.
The game uses Unreal Engine 5 to deliver 4K resolution at 60 frames each second. emptyvessel uses AWS Graviton Processors (ARM64) with UE5 to improve performance and automate development workflows.
“Our industry is at a crossroads between business and talent. At emptyvessel, we take a Developer-first approach to our direction and execution,” said Garrett Young, COO and General Manager at emptyvessel. “We understand every innovation in this industry has been led by a developer taking a risk—every new IP, every billion-dollar franchise. Gaming’s next massive hit can come from anywhere.”
Figure 1: DEFECT atmospheric squad combat environment.
Business challenge
emptyvessel, a new game development studio, set out to create an immersive multiplayer game with a globally distributed team. The studio needed an infrastructure solution that would support their vision of a fully remote workplace while maintaining high security and collaboration standards. “We recruited our talented core team of developers with a clear goal: to build a new immersive game from the ground up,” said Young. “We wanted to drive all creative decisions and work the way we’ve always wanted to – whether remote or in office.”
The distributed nature of their team presented significant technical challenges around asset management and collaboration. emptyvessel needed to implement strong security measures for their source code, art, animation, audio, and design assets while ensuring fast, reliable access to their source depot across multiple time zones. “No matter where our people work, we want them to have fast syncs to the latest change list,” explained Wei Ning, CTO at emptyvessel. “We needed 100% reliable access to our source assets around the clock.”
A critical requirement for emptyvessel was the ability to conduct multiplayer testing with up to 16 players. The studio planned to begin playtesting internally, before expanding to include investors and partners. This necessitated fast downloads and reliable build stability, as well as the flexibility to deploy game servers in multiple regions to ensure optimal gameplay experiences with low latency.
Traditional game development infrastructure, which typically assumes co-located teams with high-end hardware and fast network connections, wouldn’t meet the needs of emptyvessel. With only four engineers, two of whom were new to the industry, the studio needed to maximize development efficiency, while minimizing time spent on DevOps and infrastructure management. “We wanted to focus our programmers’ time on gameplay, animation, and performance tasks,” said Ning. “We didn’t want to overload our engineering team with build pipeline and DevOps management issues or spend hours tracking down build-breaking changes.”
These challenges called for a comprehensive cloud solution with global content distribution, automated build pipelines, low-latency networking for multiplayer testing, and cost-effective resource scaling. emptyvessel aimed to build an infrastructure that would strengthen their competitive position in the game development marketplace while supporting their vision of a distributed, developer-first studio culture.
Figure 2: Immersive shooter gameplay of dystopian environment in DEFECT.
Solution overview
Figure 3: Infrastructure stack.
“Given our studio’s belief in a cloud-native strategy as a way to accelerate and innovate game development, and the low latency requirements for running a multiplayer game globally, we decided to use AWS for both the studio and game infrastructure,” said Ning.
Building the development pipeline
emptyvessel implemented a cloud-native development pipeline using several AWS services to address their global collaboration and build management needs. The foundation of their solution begins with a Perforce database hosted on Amazon Elastic Compute Cloud (Amazon EC2), providing reliable, high-speed access to source code and development assets for their distributed team.
Implementing version control and build automation
To optimize their cloud infrastructure, emptyvessel deployed their development environment within a single Amazon Virtual Private Cloud (Amazon VPC), which hosts their version control server, build pipeline, and related microservices. This architecture maximizes internal VPC traffic to reduce costs, while AWS Global Accelerator verifies fast access for team members worldwide.
The studio implemented an event-driven architecture using Amazon EventBridge to automate their build process. When developers check in new code, EventBridge automatically triggers the build pipeline, connecting their Perforce server to multiple cloud-based builders.
“Our automated build system significantly reduced the time our engineers spent on build management,” said Ning. “By integrating build status notifications with our communication tools, we immediately know when builds pass or fail, helping us identify and resolve issues quickly.”
emptyvessel set up the builders to compile source and Unreal blueprint changes, package content, and compress new builds (for client, game server, and local host) on every single check-in throughout the day. They connected build status updates (pass or fail) directly into a company slack channel to automatically report the results of each build. Building against each change list (CL) significantly reduces the troubleshooting time needed to investigate bugs.
The studio stores build artifacts in Amazon Simple Storage Service (Amazon S3) with automated lifecycle management policies. After the engineering team validates each build, the game client is distributed to Steam, while game servers are deployed using Amazon GameLift Servers. This deployment process leverages the APIs of Amazon GameLift and Terraform modules, eliminating the need for custom deployment scripts and improving operational security.
emptyvessel further optimized their infrastructure by using the API-driven orchestration capabilities of Amazon GameLift. Moving away from traditional Secure Shell (SSH) based deployment on Amazon EC2, the studio implemented infrastructure as code using Terraform modules through AWS Cloud Control API. This modernization improved their operational efficiency and security posture.
Optimizing storage and cost management
As the project grew, emptyvessel faced increasing storage costs with build sizes reaching 10 GB. The studio implemented a cost-optimization strategy using Amazon S3 lifecycle rules. They retained regular builds for five days while using Amazon S3 replication to permanently archive milestone builds for important events (such as GDC and Gamescom). “This approach helped us balance storage costs with the need to maintain critical builds for key stakeholders,” noted Ning.
Managing game data and analytics
The studio’s data management strategy centers on Amazon DynamoDB, which provides the global availability and low latency needed for their game infrastructure. DynamoDB serves multiple critical functions, from managing build process locks to storing gameplay metrics. By routing gameplay events from Amazon EventBridge to DynamoDB, emptyvessel can analyze game performance data in near real-time.
To enhance scalability and cost efficiency, emptyvessel embraced serverless architecture. The studio implemented AWS Lambda functions for version control and release management workflows, while migrating their build pipeline to Amazon Elastic Container Service (Amazon ECS) with AWS Fargate. This serverless approach eliminated the need to manage underlying infrastructure, while activating automatic scaling based on demand.
The pipeline automatically scales to support both development and testing needs. emptyvessel applied AWS architectural patterns from web-scale industries to handle their game operations, activating efficient scaling during peak testing periods, while maintaining cost effectiveness during quieter periods.
Integration with team communication tools provides near real-time visibility into the development process. Automated notifications deliver build status updates, including change lists and stability metrics, directly to the team’s Slack channels. “This visibility helps our distributed team stay coordinated and respond quickly to any issues,” said Ning.
Automating deployment and testing
The infrastructure proved its reliability during the playtesting phase for DEFECT. Starting in January 2024, the studio conducted internal playtests two to three times a week, implementing a delta compression algorithm that reduced update sizes from 10 GB to approximately 200 MB. This optimization significantly improved the testing experience for stakeholders and potential investors.
Figure 4: DEFECT teaser image.
Looking ahead
The cloud-native game development pipeline of emptyvessel demonstrates how AWS services can support ambitious game development projects at any scale. As the studio expands the features of DEFECT, AWS will scale to accommodate increased server-side calculations, enhanced game state replication, and more complex client-side game logic. This growth includes managing larger build sizes with additional content assets, from 3D models and animations to visual effects and audio.
The studio has already implemented foundational backend services using AWS, supporting essential functions (such as player authentication, matchmaking, and game session management) for internal playtesting. Their roadmap includes expanding these services to support more robust features including cross-platform play, persistent player inventory, and comprehensive player telemetry.
“AWS provides the scalability we need for our ambitious plans,” said Ning. “As our game grows in complexity and our player base expands, we’re confident that AWS will scale to meet our needs. The ability to automatically scale resources based on demand, while maintaining cost efficiency, is crucial for a growing studio like ours.”
The combination of AWS services helps emptyvessel to focus on game development rather than infrastructure management. “Our partnership with AWS provides the worldwide build pipeline, DevOps, and live operations infrastructure we need to compete at the highest level,” said Young. “With the stability, reliability, speed, security, and scalability of AWS, we’re well-positioned to face the challenges ahead.”
Contact an AWS Representative to know how we can help accelerate your business.