This is the 3rd and final blog in our series about the NuNet private alpha. In part 1 we looked at the demo application that uses NuNet's software-hardware mesh, and part 2 went into more detail about the technical capabilities and features of this private release. In this part, we'll be looking at how NuNet's hardware-software mesh communicates with the applications running on it, and with the swarm of machines providing services, and how it joins them together to build ad hoc decentralized execution environments.
NuNet and SingularityNET strive towards similar goals in different sub-specialties: SingularityNET is building a decentralized marketplace of Artificial Intelligence agents that can work together to form a greater global intelligence; NuNet is building a decentralized computing environment of different machines that form a powerful globally decentralized computing environment. Not only do they have shared principles, there is considerable overlap in their engineering challenges. Common challenges require a common solution, which is why we’ve developed the AI-DSL (Domain-Specific Language) to consistently format the ontology across the service
The technical challenge facing NuNet is this: there is a collection of machines all over the world with spare computing resources. There are also applications with defined program graphs and hardware requirements. (These will be SingularityNET services at first, and we will later expand support to any execution environment.) NuNet’s task is to measure these and fit them together, and to do so on a dynamic, ever-changing network where machines are coming online and offline all the time.
This matchmaking process is designed with...
This article was first published on www.medium.com.
Continue reading on www.medium.com
Written by Kabir Veitas.