Sharing GPUs and Programmable Switches in a Federated Testbed with SHARY

2025-01-31

AI Summary

In federated testbeds, scarce heterogeneous hardware—such as GPUs, P4-programmable switches, and smart NICs—is shared across institutions, leading to severe resource contention, complex management, and fragmented scheduling. To address these challenges, this paper proposes SHARY, a dynamic reservation system. Its key contributions are: (1) a lightweight hardware abstraction layer that unifies diverse accelerators via protocol-agnostic abstraction; (2) the first integration of a demand-driven GPU sharing model (FIGO) and a P4-switch reservation mechanism (SUP4RNET) into a unified federated scheduling framework; and (3) tight coupling of dynamic reservation, federated identity authentication, and policy-aware coordination. Experimental evaluation demonstrates a 62% reduction in average GPU wait time and sub-200 ms P4 reservation response latency, significantly improving hardware utilization and lowering barriers to cross-institutional AI and network experimentation.

Technology Category

• Machine Learning: Hardware-aware ML
• Planning, Routing, and Scheduling: Plan Execution and Monitoring
• Multiagent Systems: Mechanism Design

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Economics, Online Markets and Human Computation: The sharing economy

Abstract

Federated testbeds enable collaborative research by providing access to diverse resources, including computing power, storage, and specialized hardware like GPUs, programmable switches and smart Network Interface Cards (NICs). Efficiently sharing these resources across federated institutions is challenging, particularly when resources are scarce and costly. GPUs are crucial for AI and machine learning research, but their high demand and expense make efficient management essential. Similarly, advanced experimentation on programmable data plane requires very expensive programmable switches (e.g., based on P4) and smart NICs. This paper introduces SHARY (SHaring Any Resource made easY), a dynamic reservation system that simplifies resource booking and management in federated environments. We show that SHARY can be adopted for heterogenous resources, thanks to an adaptation layer tailored for the specific resource considered. Indeed, it can be integrated with FIGO (Federated Infrastructure for GPU Orchestration), which enhances GPU availability through a demand-driven sharing model. By enabling real-time resource sharing and a flexible booking system, FIGO improves access to GPUs, reduces costs, and accelerates research progress. SHARY can be also integrated with SUP4RNET platform to reserve the access of P4 switches.

Problem

• Federated Testing Platform
• Resource Management
• Artificial Intelligence Research

Innovation

• SHARY
• GPU optimization
• P4 switch reservation