I am a PhD student and for a while now I'm designing and developing a distributed network protocol that enables dynamic resource allocation across heterogeneous nodes, to which I called Rank. It's designed to handle computational, network, and temporal resources in fully distributed environments without central controllers, but that could also handle a centralized environment.
Rank implements four core functions: discovery (finding paths between nodes), estimation (evaluating resource availability), allocation (reserving resources), and sharing (allowing multiple services to use the same resources).
What I think it makes it unique is its ability to operate in completely decentralized environments with heterogeneous nodes, making it particularly valuable for edge computing, cloud gaming, distributed content delivery, vehicular communications, and grid computing scenarios.
The protocol uses a bidding system where nodes evaluate their capability to fulfill resource requests on a scale from 0-1, enabling dynamic path selection based on current resource availability. I've implemented it in C++ and then also created a testing framework to validate its performance across different network topologies. This is still a work-in-progress and I am eager to publish results someday!
replies(5):