A Low-Cost Data Replication Model Based on Rsync-Syncthing for Data Loss Mitigation in MSMEs

Abstract

Micro, Small, and Medium Enterprises (MSMEs) generally store operational data on a single device without a structured backup mechanism, making them vulnerable to permanent data loss caused by hardware failure, user error, or cyberattacks. This study aimed to design and validate a low-cost data replication model based on the open-source software Rsync and Syncthing as a data loss mitigation solution that MSMEs can implement independently. The study employed a quantitative experimental approach following the Network Development Life Cycle (NDLC) framework. The tests utilized the real-world UCI Online Retail transaction dataset (531,282 rows), restructured into three MSME operational file profiles (retail, culinary, services) and evaluated under four scenarios: initial sync, incremental update, 512 KB/s network limitation, and failure simulation. The results showed that Rsync completed the initial sync in 0.035–0.195 seconds and incremental updates in 0.049–0.139 seconds; the many-small-files overhead was empirically confirmed, with the retail profile throughput 11.6 times lower than the services profile; the delta-transfer mechanism saved up to 90% of bandwidth; under a limited network, both tools converged at the bandwidth limit (±4 Mbps); and the Recovery Time Actual (RTA) reached 0.028–0.366 seconds with 100% data integrity verified by MD5 checksum.