Welcome to Cedana
Last updated
Last updated
Cedana (/ce'dana/) is Save/Migrate/Resume (SMR) for containerized CPU and GPU workloads. Our solution sits between the linux kernel and your workloads to save the full state of your container enabling you to save and resume your workloads across instances and vendors.
Our Rest API enables user to integrate this powerful capability into your applications.
We are building a global, real-time system for compute that is designed to 1) maximize compute utilization and eliminate idle resources, 2) seamlessly access all available compute, 3) increase reliabiltiy and availability through node failures, and 4) policy-based orchestration with workload-level SLAs. This changes how we allocate resources to high performance computing, numerical simulation and training and running AI models.
Take a look at the sidebar for further information. To get started with our API and running Cedana in Kubernetes, follow the sidebar!