On Energy Efficiency and Delay Minimization in Reactive Protocols in Wireless Multi-hop Networks

Abstract: In Wireless Multi-hop Networks (WMhNs), routing protocols with energy efficient and delay reduction techniques are needed to fulfill users demands. In this paper, we present Linear Programming models (LP_models) to assess and enhance reactive routing protocols. To practically examine constraints of respective LP_models over reactive protocols, we select AODV, DSR and DYMO. It is deduced from analytical simulations of LP_models in MATLAB that quick route repair reduces routing latency and optimizations of retransmission attempts results efficient energy utilization. To provide quick repair, we enhance AODV and DSR. To practically examine the efficiency of enhanced protocols in different scenarios of WMhNs, we conduct simulations using NS- 2. From simulation results, enhanced DSR and AODV achieve efficient output by optimizing routing latencies and routing load in terms of retransmission attempts.

• The paper introduces LP models to quantify energy and delay trade-offs in reactive routing protocols for WMhNs.
• The paper demonstrates protocol enhancements (AODV-LL and DSR-M) that reduce energy consumption and packet delay in dynamic network scenarios.
• The paper's results highlight the significance of rapid route repair and adaptive cache management for improving throughput in wireless multi-hop networks.

Energy Efficiency and Delay Minimization in Reactive Protocols for Wireless Multi-hop Networks

Introduction

This paper presents a comprehensive analytical and simulation-based evaluation of energy efficiency and delay minimization in reactive routing protocols for Wireless Multi-hop Networks (WMhNs). The authors develop linear programming (LP) models to formalize the trade-offs and constraints inherent in these networks, focusing on three widely used reactive protocols: AODV, DSR, and DYMO. The study further proposes enhancements to AODV and DSR, empirically validating their impact on energy consumption and end-to-end delay through extensive NS-2 simulations under varying mobility, traffic, and scalability scenarios.

Analytical Framework and LP Modeling

The core analytical contribution is the construction of LP models that encapsulate the energy and delay costs associated with the two fundamental operations in reactive protocols: Route Discovery (RD) and Route Maintenance (RM). The total energy cost for a protocol is formalized as:

$CE_{total}^{rp} = CE_{RD}^{rp} + CE_{RM}^{rp}$

where $CE_{RD}^{rp}$ and $CE_{RM}^{rp}$ represent the energy costs for RD and RM, respectively. The models account for protocol-specific mechanisms such as Expanding Ring Search (ERS), gratuitous RREPs, local link repair (LLR), and HELLO message-based link monitoring. The LP models are parameterized by network topology, node degree, and protocol-specific behaviors, enabling a comparative analysis of protocol efficiency under varying network conditions.

Protocol Enhancements

Two key protocol modifications are introduced and evaluated:

• AODV-LL: AODV is modified to utilize link layer feedback for rapid detection of link breakages, replacing periodic HELLO messages. This enables faster initiation of LLR and reduces both energy consumption and end-to-end delay.
• DSR-M: DSR is enhanced by reducing the TAP_CACHE_SIZE from 1024 to 256, expediting the removal of stale routes from the route cache. This modification aims to improve route freshness and reduce the likelihood of failed transmissions due to outdated path information.

These enhancements are designed to address the primary sources of inefficiency identified in the LP analysis: delayed route repair and excessive retransmissions due to stale or invalid routes.

Simulation Methodology

The simulation study employs the NS-2 platform with the Random Waypoint mobility model, evaluating protocol performance across a $1000 \times 1000$ m$^2$ area with 2 Mbps wireless links. Scenarios vary node mobility (2 m/s and 30 m/s), network size (10 to 100 nodes), and traffic rates (2 to 32 packets/s). Each simulation runs for 900 seconds, and performance is assessed using three metrics:

• Throughput: Successfully delivered data packets.
• Average End-to-End Delay (E2ED): Time taken for a packet to traverse from source to destination.
• Normalized Routing Load (NRL): Ratio of routing control packets to delivered data packets, serving as a proxy for energy efficiency.

Results and Analysis

Throughput

• DSR achieves high throughput at low mobility (2 m/s) due to effective RD and RM mechanisms. However, at high mobility (30 m/s), DSR's route cache becomes ineffective, leading to throughput degradation.
• AODV maintains robust throughput across all mobility and scalability scenarios, attributed to its time-based route validation and efficient RERR dissemination.
• AODV-LL outperforms standard AODV and DSR at high mobility, benefiting from rapid link break detection and repair.
• DSR-M shows improved throughput over standard DSR in dynamic scenarios due to faster cache invalidation.
• DYMO consistently underperforms in throughput, lacking supplementary mechanisms such as gratuitous RREPs or LLR.

Delay

• AODV exhibits the highest E2ED among reactive protocols, primarily due to the LLR process, which can increase path lengths during repair.
• AODV-LL significantly reduces E2ED by leveraging link layer feedback for immediate repair initiation.
• DSR maintains lower delay than AODV at low mobility but suffers increased delay at high mobility due to frequent cache misses and the overhead of searching the route cache.
• DSR-M reduces delay relative to standard DSR in high-mobility and high-traffic scenarios.
• DYMO achieves the lowest delay in low-traffic and low-mobility scenarios due to its minimalistic design but incurs higher latency as traffic and network size increase.

Energy Efficiency

• AODV-LL and DSR-M demonstrate improved energy efficiency by reducing unnecessary retransmissions and control packet overhead.
• The LP models confirm that optimizing retransmission attempts and expediting stale route deletion are critical for minimizing energy consumption in WMhNs.

Implications and Future Directions

The findings underscore the importance of rapid route repair and efficient route cache management in reactive protocols for WMhNs. The demonstrated improvements in AODV-LL and DSR-M suggest that cross-layer feedback and adaptive cache sizing are effective strategies for enhancing protocol performance under dynamic network conditions. The LP modeling approach provides a formal basis for protocol analysis and can be extended to incorporate additional constraints such as QoS requirements or heterogeneous node capabilities.

Future research directions include:

• Extending the LP models to multi-radio or multi-channel environments.
• Investigating the impact of energy harvesting and duty cycling on protocol performance.
• Integrating machine learning-based adaptive mechanisms for dynamic protocol parameter tuning.
• Evaluating protocol performance in the presence of adversarial conditions or security threats.

Conclusion

This study provides a rigorous analytical and empirical evaluation of energy efficiency and delay minimization in reactive routing protocols for WMhNs. The proposed LP models and protocol enhancements (AODV-LL and DSR-M) are shown to yield measurable improvements in throughput, delay, and energy consumption, particularly under high mobility and scalability scenarios. The results highlight the value of rapid route repair and adaptive cache management, offering actionable insights for the design of next-generation WMhN routing protocols.