An Energy-Aware Recovery-Line Optimization Algorithm for Improving Reliability and Recovery Efficiency in Mobile Computing Systems
DOI: 10.66838/ma.246
Dr. Ajay Sharma Research Guide (Computer science), Department of Computer Science and Engineering, Shri Venkateshwara University, Gajraula, UP, India.
|
Abstract Energy consumption and recovery efficiency are two of the most pressing concerns in the design of fault-tolerant mobile computing systems. Conventional recovery-line management protocols treat these two objectives independently: energy-oblivious protocols maximize consistency at the cost of excessive checkpoint power draw, while energy-aware protocols reduce battery consumption at the cost of increased rollback distance or weakened recovery guarantees. This paper proposes EARLA—the Energy-Aware Recovery-Line Optimization Algorithm—a novel algorithm that unifies energy management and recovery-line optimization into a single adaptive decision framework. EARLA models each mobile host as a four-state energy machine (Active, Moderate, Low, Critical) and dynamically selects the optimal checkpointing action from a repertoire of four operations—full checkpointing, mutable checkpoint transfer, log-only deferral, and complete deferral—based on the current energy state, estimated mobility rate, and network quality score. A recovery-line index (RLI) maintained across three storage tiers ensures that the recovery line advances at every epoch without global coordination. Formal analysis establishes that EARLA is domino-effect free under crash fault assumptions. Simulation experiments over 1000 Monte Carlo runs with 20 mobile hosts demonstrate that EARLA achieves an energy consumption of 22.7 mJ per checkpoint interval (56.4% lower than coordinated checkpointing), a mean time to failure (MTTF) ratio of 1.15× (15% better reliability than the coordinated baseline), a recovery time of 142 ms (70.4% faster than coordinated checkpointing), and a normalized rollback distance of 0.26× with zero domino-effect risk. These results establish EARLA as the most energy-efficient and recovery-efficient protocol among all evaluated alternatives.
Keywords
Energy-aware checkpointing, recovery-line optimization, mobile computing, fault tolerance, MTTF, rollback reduction, adaptive algorithm, battery-aware scheduling., |