Answers are grounded in this paper and cite their sources.
In hospital Internet of Medical Things (IoMT) deployments, Edge gateways process routine sensor data and occasional anomalous windows that need urgent analysis. On a congested gateway the urgent analysis competes with routine work, whereas forwarding it to the hospital Fog node adds network delay and traffic and may overload it. We propose CriticalityAware, a rule executed by each gateway: a critical tuple is forwarded to the Fog node only if the gateway utilization over the last 100 ms reaches a threshold and the utilization advertised by the Fog node is below a second threshold; otherwise it is processed locally. Using an extended iFogSim, we compared the rule with Edge-local processing and unconditional forwarding in a four-tier topology. At load factors of 1.0 and 1.2 and with heterogeneous gateways, CriticalityAware reduced the mean latency of critical tuples by 5–12% relative to Edge-local processing and, at those loads, by 13–38% relative to unconditional forwarding, which generated 1.5–3.3 times more traffic. At the highest load, Edge-local processing was faster for critical tuples but doubled the latency of routine data, and unconditional forwarding saturated the Fog node. Absolute gains are a few milliseconds; the rule’s main benefit is robustness, which depends on the Fog-side threshold.
No takes yet. Share an insight, caveat, or question.
Torres et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: