Table 3 Testbed Description.
Component | Details |
|---|---|
Hardware | 5 compute nodes, each with 16 vCPUs, 64 GB RAM, and 2Â TB storage (SSD) |
Client Machines | 16 GB RAM to simulate healthcare providers accessing data |
Network | 1 Gbps LAN, star topology, TLS/SSL encryption, simulated latency |
Dataset Sizes | 500,000 medical images (1.5Â TB), 100 GB metadata (NIH Chest X-ray, TCIA) |
Transaction Data | 284,807 credit card transactions, 31 anonymized features (~ 100 MB) |
Workload Generator | Poisson distribution (100 requests/min), batch requests |
Arrival Processes | Poisson process for request simulation, batch data requests |
Caching | LRU caching, 10 GB capacity, cache hit rate tracking |
Metrics Tracked | Authentication time, access delay, anomaly detection accuracy, throughput, cache hit rate |
Measurement Tools | Prometheus (monitoring), Grafana (visualization), Python (logging & analysis) |
Statistical Tests | Paired t-tests for comparisons, 95% Confidence Intervals (CIs), 5–10 runs |
Experiment Runs | 5–10 runs per configuration for consistency in results |
