X-DNet: Energy-Efficient Distributed and In-Network Computing via Cross-Layer ApproXimation of Applications and Accelerators Software Campus (BMBF)
Developing energy-efficient and real-time solutions are from the main challenges in the era of 5G/6G, especially considering the ever-increasing complexity of computational algorithms in stream processing and AI-based applications. To address these concerns, computing is envisioned to become distributed and/or performed on-the-fly, while data is transmitted through the network elements (dubbed as In-Network Computing). In this context, the goal of X-DNet project is to improve performance- and energy-efficiency, through a HW/SW co-design approach. To achieve this end, we first reduce the complexity of applications using various ‘approximate computing’ techniques. Afterwards, we design different accelerator configurations, to be deployed in various network elements within the edge-to-cloud continuum.
