GTO Wizard Migrates from On-Premises to AWS

Intro

GTO Wizard is the #1 application for serious poker players, offering a suite of powerful tools to study, practice and analyze poker strategy. Their platform is built on the world’s most advanced GTO (Game Theory Optimal) solver, a high-intensity compute engine that requires massive processing power. Initially housed in an on-premises data center, their infrastructure faced significant challenges in scaling, cost management and operational efficiency as the platform’s popularity surged.

To build a foundation for future growth, GTO Wizard decided to migrate to AWS. GOStack partnered with Revolgy to design the new AWS infrastructure, focusing on best practices in cloud architecture, DevOps, GitOps and FinOps to create a platform that was not only powerful but also scalable and cost-efficient.

The Challenge

The GTO Wizard platform is unique in its reliance on high-intensity, on-demand compute to run its sophisticated AI solver. Their on-premises environment presented several critical limitations:

Scaling Bottlenecks: The physical constraints of the data center made it difficult and slow to scale compute resources to meet the fluctuating demands of a global user base. This led to performance issues and limited their ability to grow.

High Operational Costs: Managing a large, power-hungry on-premises environment incurred significant capital and operational expenditure, from hardware procurement and maintenance to power and cooling.

Slow Delivery Cycles: The lack of a modern DevOps culture and automation meant that deploying new features and updates was a slow, manual process, hindering their ability to innovate and respond to user feedback.

Lack of Cost Visibility: Without the granular cost allocation and monitoring tools available in the cloud, it was difficult to track infrastructure spend and optimise for efficiency.

Our Solution

Working in close partnership with Revolgy, GOStack designed a new, modern AWS infrastructure from the ground up, tailored to the unique demands of the GTO Wizard platform.

Migration to Amazon EKS: The core of the new platform is Amazon EKS, providing a scalable and resilient environment for the containerized GTO Wizard application. This allows for the efficient management of both the core application services and the high-intensity compute workloads required by the GTO solver.

Infrastructure as Code and GitOps: The entire AWS environment was defined using Terraform, establishing a single source of truth for the infrastructure. Combined with a GitOps workflow using ArgoCD, this enables a fully automated, auditable and repeatable process for provisioning and managing the platform. The GTO Wizard team can now manage their infrastructure with the same pull-request-based workflow they use for their application code.

High-Performance Data Layer: To handle the massive datasets generated by the solver and user analysis, we implemented a robust data platform. ClickHouse was chosen for its exceptional performance in real-time analytics, supplemented by Amazon RDS for structured data and Amazon DynamoDB for low-latency key-value storage.

FinOps for Cost Optimization: A key focus of the project was implementing FinOps best practices. By leveraging AWS cost management tools, right-sizing resources and using services like AWS Lambda for event-driven tasks, we established a culture of cost accountability. This ensures that the platform runs as efficiently as possible, with clear visibility into spend.

Security and Performance: Amazon CloudFront was deployed as a global CDN to deliver low-latency access to users worldwide, while AWS WAF protects the application from common web-based threats.

Results and Benefits

The migration to AWS, designed by GOStack and Revolgy, has fundamentally transformed GTO Wizard’s technical capabilities.

• Improved Developer Productivity: The self-service platform and automated GitOps workflow have eliminated delivery bottlenecks, allowing the development team to deploy new features faster and more frequently.

• Enhanced System Stability: The modern, scalable architecture on AWS has significantly improved the stability and performance of the platform, even under heavy load from the high-intensity compute engine.

• Cost Optimization: Through FinOps practices and the use of efficient AWS services, GTO Wizard has achieved greater cost-effectiveness and a clear, predictable spending model for their infrastructure.

• Scalability for Growth: The platform is now built to scale, ready to support GTO Wizard’s continued growth and their expansion into new products and services.