From Devices and Desktops to the Cloud: Massive Scale Hybrid Computing

About the Session

The Distributed Compute Protocol (DCP) is a client-side framework built on a web-standard execution model. Developed by Distributive, DCP harmonizes browser, standalone, and embedded environments to deliver a "write-once, run-everywhere" distributed compute fabric. It seamlessly aggregates cycles across a massive range of hardware:

• Edge & IoT: Webpages, smart screens, and school lab machines.

• Enterprise: On-premise workstations and idle server capacity.

• Hyperscale: Elastic cloud instances.

For organizations managing complex workloads, DCP offers built-in policy controls that allow public networks and private on-prem environments to be pooled with cloud resources as a single, logical system. This enables a unique hybrid strategy: reducing the Total Cost of Ownership (TCO) by harvesting existing "sunk-cost" hardware while maintaining the ability to burst into the cloud for massive scale.

In this session, Dan Desjardins will demonstrate the true flexibility of DCP by forming a live, heterogeneous cluster composed of smart devices, local hardware, and GCP servers.

The Challenge:

• Infrastructure: Combining highly unconventional edge devices with high-performance GCP compute resources.

• Workload: Deploying a real-world Monte Carlo simulation for human-resources modeling.

• Execution: Watch the workload distribute and compute in real-time across this mixed infrastructure.

As we look toward "AI for Science" and increasingly complex simulation requirements, the ability to orchestrate computational capacity across every available chip is a force multiplier. Learn how universities, hospitals, and government labs are already using DCP to build platforms that complement their core HPC investments with elastic cloud capacity.

Speaker