Evaluating the impact of static and dynamic network bandwidth on checkpointing-based fault tolerance in distributed

Distributed systems form the backbone of modern computing infrastructure and demand robust fault tolerance to ensure continuous operation amidst failures. This study evaluates how static and dynamic network bandwidth conditions affect the performance of checkpointing-based fault-tolerant mechanisms. Unlike traditional approaches that assume constant bandwidth, our research simulates realistic, fluctuating bandwidth scenarios. We propose an enhanced coordinated check-pointing algorithm that adapts to network variability and evaluates it using Python and Message Passing Interface (MPI) across 1,000 processes and 1,000 test cases. Our analysis, based on over 2 million time measurements, reveals that variable bandwidth introduces measurable latency, with fluctuations ranging from -0.41661-0.333473 microseconds in the first 50 cycles. These findings highlight the need for adaptive fault tolerance strategies and provide valuable insights for building resilient distributed systems in heterogeneous network environments.