Abstract
Due to the near-zero idle power consumption characteristic of non-volatile random access memory (NVRAM), NVRAM has gained popularity as a great alternative to volatile RAM on energy-constraint and tiny embedded systems that run on intermittent power. In particular, multi-level-cell (MLC) NVRAM can further reduce the energy consumption of data writes by alternating data write modes with different retention periods. Nevertheless, due to the unstable power supply of intermittent systems, data on NVRAM could become obsolete and result in distortion of reality as time goes by. As intermittent systems typically involve a large number of sensors and numerous transmitters to collect environmental data, data freshness is vital for accurate scientific study. To ensure the data freshness, this paper proposes a minimizing average system age-of-information (AoI) scheme, abbreviated as MASA scheme, for MLC-NVRAM-based intermittent systems. The proposed policy harmonizes data write modes according to data update intervals of sensors with the goal of minimizing AoI and energy consumption. The experimental results indicate that the proposed policy achieves a better average AoI than the existing schemes.