Documentation

Introduction

Expanos provides EC2-equivalent machines designed to integrate seamlessly into AWS EKS (Elastic Kubernetes Service) clusters, functioning as worker nodes while delivering competitive pricing.

The economics of public cloud often work against teams with high computational demands. Providers like AWS charge a premium — up to 10x the cost of equivalent hardware in a private data center. However, running a private data center isn’t feasible for most teams due to its complexity and scale requirements, leaving this option only to enterprises with sizeable devops teams.

Expanos solves this by extending your existing AWS EKS infrastructure with physical, single-tenant servers hosted in trusted data centers. This approach retains the scalability, security, and managed services of AWS while offloading compute-intensive workloads to cost-efficient, high-performance machines.

Unlike cloud providers, Expanos does not require long-term commitments to save costs. You can request nodes on a monthly basis. We recommend you to reserve just enough to handle baseline workloads, and let AWS auto-scale as needed for peak demand.

Why is Expanos so cost-effective?

• Pre-paid servers: Thanks to our model, we know the demand for our servers ahead of time. We do not need to keep spare capacity idling in case of peak demand. Expanos only expands existing clusters, leaving autoscaling to AWS.
• Negotiated prices: We pre-pay for servers at discounted rates by partnering with trusted data centers. This allows us to offer our customers the same price as if they were buying hardware from the data centers directly themselves.
• Picking from multiple data centers: We constantly compare prices across different data centers to secure the most cost-effective resources, especially for GPU workloads.
• Purpose-built infrastructure: Since we are not competing with cloud providers, but building on top of their infrastructure, we do not have to rebuild the entire cloud. We focus solely on Kubernetes workloads, leaving the rest such as managed databases, storage, load balancing, autoscaling, and networking to AWS.

Where are the servers located?

Expanos leverages a network of vetted data centers worldwide. We continually evaluate providers to ensure competitive pricing and optimal performance, particularly for GPU-intensive workloads.

Does AWS allow this?

Yes, Expanos operates within the constraints of AWS policies and augments your existing infrastructure rather than replacing it.

We are working on an official partnership certification and you will soon be able to find Expanos in AWS Marketplace and pay for it using your existing credits.

Recommended use cases

Expanos is ideal for:

• Compute-intensive workloads such as machine learning and large-scale data processing
• CI/CD pipelines
• Internal tools

We do not recommend Expanos for workloads that require low latency and high availability such as databases or file servers. While Expanos machines are generally stable and perform well, they were designed to be used as Kubernetes worker nodes often running in different regions.

Additionally, we do not recommend Expanos for high-bandwidth applications such as video streaming or transcoding. Unfortunately cloud providers charge excessive rates for egress traffic, which negates the benefit of cheap compute.