How Does Kinesis Network Work?

Kinesis connects datacenters, clouds, and customer hardware into one compute grid. Like an electricity grid, it brings multiple sources behind a shared service. You choose the requirements; the platform coordinates the supply.

Choose where your workload runs

Set hardware, location, and placement requirements for each app. Use Kinesis-managed capacity, connect your own infrastructure, or combine them. The right choice depends on performance, cost, and data requirements.

Deploy your app

1. Import a container, connect a repo, or choose an app in the Portal.
2. Set runtime and hardware requirements, including CPU, RAM, and GPU capacity.
3. Start the deployment. Kinesis places it on suitable compute and monitors its health.

The platform handles scaling and recovery within your configured constraints. You can inspect logs, metrics, and deployment status from the same console.

Connect to your workload

Load balancing routes traffic to healthy instances. HTTP and TCP endpoints give your applications access to services running on the grid.

Understand usage and cost

Serverless pricing meters compute, memory, storage, and network usage. Dedicated compute uses hourly rates. See the pricing page for configurations and billing terms.

Usage is visible per application, so you can identify which workloads drive cost and where resources are underused.

Connect capacity you already own

Bring cloud reservations, on-prem servers, or other supported hardware into the grid. Use it for your workloads or offer spare capacity to customers. See Grid Your Compute and provider options for the two paths.