GPUs Kubernetes Serverless Platform Engineering Machine Learning Artificial Intelligence AI ML

Simplify GPUs usage across any Kubernetes Cluster

Clastix announces a new Serverless GPU solution developed in partnership with Seeweb, Italy's leading provider of cloud infrastructures for Artificial Intelligence.

Monday, March 25, 2024 Adriano Pezzuto

Revolutionizing GPUs usage for AI

Leveraging Seeweb's robust computational infrastructure, we're excited to announce a new service to transform how AI teams access and utilize GPUs for their projects. Designed to challenge traditional GPUs rental model, the new service enables on-demand GPUs access anytime, anywhere, and at the most competitive price.

It allows AI teams to bypass the upfront costs associated with on-premises hardware procurement and avoid the rigid, long-term commitments typically demanded by traditional GPU providers.

Traditionally, when users opt to rent a GPU from a cloud vendor, it’d likely need to:

  • Select a GPU, based on model performance requirements 

  • Launch a remote machine hosting the GPU 

  • Login in that machine, install and set up all the necessary software, e.g., GPU drivers, libraries, container runtime, orchestrator, etc. 

  • Containerize the model on local machine

  • Transfer the container to the remote machine, 

  • Install serving software depending on model and deployment requirements

  • Configure the serving software to load the model

  • Finally, run the model 

Instead, with serverless GPU, now users can effortlessly access any remote GPUs directly from their local Kubernetes cluster, integrating seamlessly within their environment and with the tools they already know and trust. Just installing a lightweight agent on local Kubernetes and a virtual node presenting remote GPUs is immediately available for scheduling.

The service stands out with rapid spin-up times and responsive auto-scaling capabilities, eliminating the inefficiencies of the traditional way. AI teams can now flexibly scale their computational resources, spin up GPUs per request, scale down once done, and pay only for the actual consumption. This solution is designed to meet the computational needs of AI workloads from the smallest to the largest scales.

Seeweb's power meets Clastix's kubernetes expertise

At the heart of this service lies a synergy between Seeweb's computational capacity, specially optimized for AI, and Clastix's innovative approach to Kubernetes. This partnership embodies our shared vision of making GPU technology accessible and affordable for AI projects of all sizes. Seeweb is a leader in cloud computing infrastructure in Italy, offering a wide range of services, from web hosting to cloud solutions. Committed to reliability and performance, Seeweb enables businesses to leverage the power of the cloud for growth and innovation.

Getting Ready

Stay tuned for further updates, and prepare to unlock unprecedented flexibility and efficiency in your AI projects.