Hamming Tree: The Case for Energy-Aware Indexing for NVMs

Abstract

Non-volatile memory (NVM) technologies are widely adopted in data storage solutions and battery-powered mobile and IoT devices. Wear-out and energy efficiency are two vital challenges facing the use of NVM. In Hamming Tree, we propose a software-level memory-aware solution that picks the memory segment of where a write operation is applied judiciously to minimize bit flipping. It has been shown that reducing bit flips leads to reducing energy consumption and improving write endurance. We performed real evaluations on an Optane memory device that show that Hamming Tree can achieve up to 67.8% reduction in energy consumption.