System Support for Environmentally Sustainable Computing in Data Centers

TL;DR

The paper addresses the challenge of sustaining QoS in data centers while reducing both operational and embodied carbon under renewable-energy variability and hardware recycling. It proposes a three-part system comprising Amoeba, a FeFET-based reconfigurable PIM accelerator; FRAC, a fraction NAND flash scheme with adjustable V_th states to extend recycled flash lifetime; and ESE, an environmental sustainability estimator that combines a hardware task partitioning estimator, an energy model with $E_{ope}$ and $E_{emb}$, and an energy-source predictor. The approach targets energy-efficient memory-intensive workloads, lifetime extension for recycled storage, and accurate energy accounting to drive sustainable incentives. Preliminary results indicate carbon-minimization advantages for Amoeba, measurable lifetime-extension potential for FRAC, and plausible energy forecasting for ESE, suggesting meaningful practical impact for scalable, low-carbon computing in data centers.

Abstract

Modern data centers suffer from a growing carbon footprint due to insufficient support for environmental sustainability. While hardware accelerators and renewable energy have been utilized to enhance sustainability, addressing Quality of Service (QoS) degradation caused by renewable energy supply and hardware recycling remains challenging: (1) prior accelerators exhibit significant carbon footprints due to limited reconfigurability and inability to adapt to renewable energy fluctuations; (2) integrating recycled NAND flash chips in data centers poses challenges due to their short lifetime, increasing energy consumption; (3) the absence of a sustainability estimator impedes data centers and users in evaluating and improving their environmental impact. This study aims to improve system support for environmentally sustainable data centers by proposing a reconfigurable hardware accelerator for intensive computing primitives and developing a fractional NAND flash cell to extend the lifetime of recycled flash chips while supporting graceful capacity degradation. We also introduce a sustainability estimator to evaluate user task energy consumption and promote sustainable practices. We present our preliminary results and recognize this as an ongoing initiative with significant potential to advance environmentally sustainable computing in data centers and stimulate further exploration in this critical research domain.

Figures (7)

• Figure 1: The Amoeba architecture.
• Figure 2: A Fraction NAND flash cell (FRAC)
• Figure 3: A fully nonvolatile accelerator.
• Figure 4: An environmental sustainability estimator (ESE).
• Figure 5: Carbon footprints and progresses under energy fluctuations compared to prior designs.