Integrating IoT Sensors and Smart Automation into Traditional Irrigation Systems (Stepwells, Johads, Ahar-Pyne): Water Conservation and SDG 6

Krunal R. Parmar — 2026-10-28

India’s traditional water systems (stepwells, johads, and the ahar-pyne network of south Bihar) store a brief monsoon and stretch it across a long dry period. They feed groundwater, keep evaporation low, and move water by gravity with no grid power, yet they run on community habit and visual judgement rather than measurement, which makes their conservation value hard to verify or scale. Inexpensive IoT sensing, low-power radio links, edge computing, and automated actuators are now practical for water management. This study reviews how the two can be joined in support of Sustainable Development Goal 6. It outlines the three systems and their hydrology, works through the water balance of a johad using assumed inputs, surveys suitable sensors, links, analytics and actuators, and proposes a community-centred integration in which traditional knowledge supplies the operating rules while IoT supplies measurement, checking and timely alerts. Cost, upkeep, heritage limits, connectivity, governance and data ownership are discussed. The review concludes that IoT should support traditional systems, not replace them, and that quantitative field evidence for IoT-assisted traditional structures is still scarce.

Latency-Aware Edge–Cloud Task Placement for Reliable Sensor Stream Analytics Under Variable Network Conditions

Aashi Singh Bhadouria — 2026-10-29

Background: Sensor analytics increasingly spans devices, edge gateways, and the cloud. Sending every task to the cloud exposes it to wide-area delay and congestion, whereas running everything at the edge overloads limited processors, and both choices cause missed deadlines.

Objectives: This study proposes Latency-Aware Deadline-driven Placement (LADP), an interpretable policy for deciding where each sensor-analytics task should run, and quantifies its benefits and costs.

Methodology: For each arriving task, LADP predicts completion times at the edge and in the cloud from queue backlogs, exponentially weighted estimates of round-trip time and bandwidth, and profiled compute demand; it restricts the choice to targets predicted to meet the deadline, selects among them by a latency-plus-bandwidth cost, and serves both the edge processor and the uplink in earliest-deadline-first order. An event-driven simulator with four task classes, bursty arrivals, a four-core edge server, a shared uplink, and a wide-area network with congestion episodes was used for 1,130 paired runs.

Results: LADP had the lowest deadline-miss ratio at every load, round-trip time, and bandwidth tested. At 10 tasks per second, it missed 11.0% of deadlines, versus 16.3% for size-threshold offloading and 65.1% for cloud-only execution, and reduced 95th-percentile latency by 44% and halved uplink traffic relative to threshold offloading.

Conclusions: Placement and deadline-aware scheduling should be designed together; better network estimation is the main avenue for improvement.

Journal of Sensor and Cloud Computing (e-ISSN: 3048-9199)

Integrating IoT Sensors and Smart Automation into Traditional Irrigation Systems (Stepwells, Johads, Ahar-Pyne): Water Conservation and SDG 6

Latency-Aware Edge–Cloud Task Placement for Reliable Sensor Stream Analytics Under Variable Network Conditions