Journal of Networking Technology Oothongsap, P. Mingkhwan A. 17 2 2026 https://doi.org/10.6025/jnt/2026/17/2/83-94 https://www.dline.info/jnt/fulltext/v17n2/jntv17n2_3.pdf This study presents a methodological framework for analyzing long-range dependence (LRD) in IoT blockchain time series through fractal analysis. Utilizing a synthetically generated Fractional Gaussian Noise (FGN) dataset that emulates high speed network traffic, the research employs the Hurst exponent as a primary metric to quantify temporal persistence and scaling behavior. Two classical yet robust techniques Rescaled Range (R/S) analysis and Detrended Fluctuation Analysis (DFA) are systematically applied to estimate the Hurst exponent across multiple temporal scales. The results demonstrate strong power law scaling, with linear log log relationships confirming scale invariant dynamics. Quantitative estimates yield Hurst exponents of approximately 0.91 (R/S) and 0.88 (DFA), both significantly exceeding the 0.5 threshold for random walks, thereby indicating strong persistent behavior and long term memory. Corresponding fractal dimension values (1.09 and 1.12) reveal moderate smoothness combined with structured irregularity, highlighting the intrinsic complexity of the underlying processes. The high consistency between R/S and DFA outcomes validates the framework's reliability, while DFA's detrending mechanism ensures robustness to nonstationarity. These findings carry significant implications for IoT blockchain ecosystems, particularly in predictive modeling, multi-scale anomaly detection, and adaptive resource allocation. By capturing selfsimilar traffic patterns, the framework enhances system resilience against congestion and stealthy cyber threats. Future research will extend this univariate approach to multivariate network data and explore realtime fractal monitoring for edge-computing environments.