Efficient Distributed DNNs in the Mobile-edge-cloud Continuum

TL;DR

This paper introduces RightTrain, a joint decision-making framework for optimizing data, model, and resource matching in distributed deep neural network training across mobile, edge, and cloud nodes, reducing energy use while maintaining quality.

Contribution

It presents a novel formulation and a polynomial-time solution, leveraging an expanded-graph and Steiner tree, with provable near-optimality and significant performance improvements.

Findings

Runs in polynomial time with a competitive ratio of 2(1+ε)

Outperforms state-of-the-art solutions by over 50%

Validated through real-world implementation

Abstract

In the mobile-edge-cloud continuum, a plethora of heterogeneous data sources and computation-capable nodes are available. Such nodes can cooperate to perform a distributed learning task, aided by a learning controller (often located at the network edge). The controller is required to make decisions concerning (i) data selection, i.e., which data sources to use; (ii) model selection, i.e., which machine learning model to adopt, and (iii) matching between the layers of the model and the available physical nodes. All these decisions influence each other, to a significant extent and often in counter-intuitive ways. In this paper, we formulate a problem addressing all of the above aspects and present a solution concept called RightTrain, aiming at making the aforementioned decisions in a joint manner, minimizing energy consumption subject to learning quality and latency constraints. RightTrain leverages an expanded-graph representation of the system and a delay-aware Steiner tree to obtain a provably near-optimal solution while keeping the time complexity low. Specifically, it runs in polynomial time and its decisions exhibit a competitive ratio of $2(1+\epsilon)$, outperforming state-of-the-art solutions by over 50%. Our approach is also validated through a real-world implementation.