VALDI pioneers the ability to run isolated Docker containers on a distributed cloud computing environment. This workflow is especially ideal for long-running tasks such as Large Language Model (LLM) training, hosting APIs for Machine Learning inference, data mining, and much more.
In many cases — especially during the innovation phase — it’s easier and more natural for researchers and developers to work on, interact with, and debug their code directly on the high-performance computing (HPC) workstations where they will ultimately run their applications.
To this end, VALDI will soon launch Direct SSH Access to network devices. Users will be able to “ssh into” a completely sandboxed environment on any available network device, giving them the ability to mimic the UX of more traditional virtual cloud servers, like AWS EC2, but at significantly lower cost.
By leveraging kernel features such as namespaces, cgroups, and layered filesystems, Compute Providers’ host operating systems are inaccessible by users working with the isolated VALDI virtual environment on the machine.
With this update, VALDI can be more easily integrated into existing R&D workflows and ultimately be leveraged in more creative and impactful ways.