Case Studies

Hathora Services 6 Continents with a Lean Team and Hybrid Cloud Ops

Edge Bare Metal Hybrid Retail Gaming Location Global
hathora case study

Founded in 2022, Hathora has quickly established itself as a specialized platform-as-a-service provider for multiplayer game developers. They deliver specialized infrastructures that enable studios of all sizes to deploy high-performance game servers worldwide. While popular, platforms like Amazon EKS are not generally cost-effective for businesses in the gaming industry, where high-performance, low-latency servers are critical. Instead, large studios like EA, Blizzard, and Valve often turn in-house, developing proprietary systems of their own. Hathora aims to eliminate the need for studios to reinvent server infrastructure by providing a purpose-built platform tailored to this industry’s unique requirements, giving the historically underserved gaming industry the tailored solution it needs.

Challenge

  • Cost effective compute
  • High-performace, low-latency servers

Environment

  • 30,000 cores in compute clusters
  • Clusters distributed across 14 regions
  • Managed by 6 engineers
  • Omni management
Impact

Impact

  • Savings on licensing and cloud usage costs
  • Unified management
  • Ability to onboard worker machines from any platform
  • No vendor lock-in
Question

Why Sidero and Omni

  • Efficient platform as a service for game developers
  • Ability to support multiple bare metal providers across multiple regions
Challenge

Six Engineers Re-imagine Infrastructure to Slash Costs for Customers

“From a cost perspective, there’s just a ton of bandwidth involved with gaming because you have 100 updates every second sent to every connected player. And we all know how much the major cloud providers like to charge for bandwidth. That’s what motivated our shift to bare metal.”

Harsh Pandey, CTO and co-founder, Hathora

Hathora discovered that building a platform on any major cloud provider for orchestration was cost-prohibitive and not efficient at scale. The enormous infrastructure costs of the compute resources to serve a global network of players are one of–if not the–biggest expenses for game developers. Hathora also needed a secure-by-default system that could support both a lean team and their large customers. This required a new strategy for building a cost-effective infrastructure. To achieve this, Hathora chose to reduce their dependency on cloud and leverage bare metal to create a fully hybrid environment. They needed an orchestration solution to seamlessly manage bare metal and cloud infrastructures to scale across multiple cloud providers and regions.
Solution

True hybrid cloud operations with added cost savings, courtesy of Omni and Talos

Hathora evaluated several options, including EKS Anywhere and Rancher, and found they didn’t have the full functionality needed, and the cost was prohibitively high. Then, they discovered Talos Linux. With Talos Linux able to run on any node, Hathora realized their compute resources would be consistent across their entire infrastructure, regardless of the underlying provider. As they implemented Talos Linux across all of their nodes, they realized a unified foundation for their entire infrastructure, spanning both cloud and bare metal.

Even with Talos Linux providing consistent APIs and management across all infrastructure providers, Hathora needed to be able to efficiently manage, operate and scale dozens of Kubernetes clusters and thousands of nodes, simultaneously. They added Omni to centrally create, upgrade, and manage all of their clusters with an enterprise-grade solution, ensuring simplified and secured Kubernetes across bare metal, data center, hybrid, and edge.

With this flexible and scalable infrastructure, Hathora is able to work with multiple bare metal providers alongside GCP and AWS, with plans to onboard more providers in the future. They now support an 80/20 split between bare metal and cloud resources, using bare metal for baseline capacity and turning to cloud resources only during peak hours to reduce costs.

Results

Cost-Optimized Infrastructure with Enterprise-Grade Security

“We have about six engineers on a team of 11 in total, and we service 14 regions. A lot of people are very surprised when they hear this. We wouldn’t be able to achieve this without the unified management that Talos Linux and Omni provide.”

Harsh Pandey, CTO and co-founder, Hathora

With Omni and Talos Linux, Hathora successfully built a hybrid infrastructure for the efficient management of Kubernetes at scale, supporting both their rapid business growth and their customers’ needs without having to scale their team. The 80/20 split between bare metal and cloud has enabled them to significantly reduce infrastructure costs without sacrificing performance.

Additionally, they have been able to integrate new providers to easily scale, and have the flexibility to use multiple compute environments while avoiding lock-in. The reliable security and the removal of SSH allows the team to ensure security for their customers at scale.

Hathora’s deployment of Talos Linux and Omni demonstrates how purpose-built infrastructure can help specialized workloads achieve the perfect balance of performance, cost, and operational efficiency. By combining bare metal economics with cloud elasticity, Hathora is able to orchestrate Kubernetes reliably and strategically.

Thanks for reading!

There's more to discover.

Simply click through for the next article.

Previous Next