A Hop-by-Hop Congestion-Aware Routing Protocol for Heterogeneous Mobile Ad-hoc Networks

In Heterogeneous mobile ad hoc networks (MANETs) congestion occurs with limited resources. Due to the shared wireless channel and dynamic topology, packet transmissions suffer from interference and fading. In heterogeneous ad hoc networks, throughput via a given route is depending on the minimum data rate of all its links. In a route of links with various data rates, if a high data rate node forwards more traffic to a low data rate node, there is a chance of congestion, which leads to long queuing delays in such routes. Since hop count is used as a routing metric in traditional routing, it do not adapt well to mobile nodes. A congestion-aware routing metric for MANETs should incorporate transmission capability, reliability, and congestion around a link. In this paper, we propose to develop a hop-by-hop congestion aware routing protocol which employs a combined weight value as a routing metric, based on the data rate, queuing delay, link quality and MAC overhead. Among the discovered routes, the route with minimum cost index is selected, which is based on the node weight of all the in-network nodes. Simulation results prove that our proposed routing protocol attains high throughput and packet delivery ratio, by reducing the packet drop and delay.

💡 Research Summary

The paper addresses the problem of congestion in heterogeneous mobile ad‑hoc networks (MANETs), where nodes differ in transmission power, data rate, processing capability, and reliability. Traditional MANET routing protocols, such as AODV, rely on hop count as the sole metric, which fails to capture the asymmetric capabilities of heterogeneous nodes and often leads to bottlenecks when high‑rate nodes forward traffic to low‑rate neighbors. The authors propose a hop‑by‑hop congestion‑aware routing protocol (CARP) that integrates four key link‑level parameters into a single composite weight: (1) data‑rate, (2) buffer queuing delay, (3) link quality (derived from received signal strength), and (4) MAC overhead (time spent in RTS/CTS/ACK exchanges and contention).

The composite metric, called node weight (NW), is defined as the product of the four parameters (NW = qL × MAC_OH × Rate × QueueDelay). During route discovery, each intermediate node measures the four parameters locally, computes its NW for the outgoing link, and adds this value to the cumulative weight carried in the Route Request (RREQ) packet. When the RREQ reaches the destination, the total accumulated weight is included in the Route Reply (RREP) that traverses back to the source. Each node on the reverse path updates its own cost based on the received total weight, enabling the source to compare multiple disjoint paths (obtained via an AOMDV‑style multipath discovery) and select the one with the minimum total NW.

To further guard against inaccurate delay estimates that can arise from route‑request timing, the protocol introduces a “dummy‑RREP” phase. After a candidate path is identified, the source sends a short burst of dummy data packets (identical in size, priority, and data rate to real traffic) along that path. The destination measures the average end‑to‑end delay of these dummy packets and returns the value to the source in a special RREP. If the measured delay satisfies the application’s bound, the path is accepted; otherwise, the source switches to the next best candidate. This proactive verification reduces the likelihood of selecting a path that would later experience excessive queuing or link failures.

The authors also propose a simple effective data‑rate estimation formula (Rate = PacketSize / ChannelDelay) that captures the reduction in usable bandwidth caused by congestion‑induced retransmissions and queuing.

Simulation experiments were conducted using an IEEE 802.11b physical layer with heterogeneous nodes (different transmission powers and data rates). The proposed CARP was compared against AODV, CARM (a previous congestion‑aware metric using data‑rate, MAC overhead, and buffer delay), and CRP (a congestion‑adaptive routing protocol). Performance metrics included throughput, packet delivery ratio (PDR), average end‑to‑end delay, and packet loss ratio under varying traffic loads. Results showed that CARP achieved a PDR improvement of more than 15 % over the baseline protocols, reduced average delay by about 30 % in high‑load scenarios, and increased overall network throughput by roughly 20 % by preferentially selecting routes whose bottleneck link had the highest effective data rate. Moreover, the proactive dummy‑packet verification cut down packet loss caused by sudden link degradation.

In conclusion, the paper demonstrates that a routing metric which jointly considers physical‑layer data rate, MAC‑layer contention, link‑quality, and queuing delay can effectively mitigate congestion in heterogeneous MANETs. The hop‑by‑hop accumulation of this composite weight enables dynamic, cost‑aware path selection without requiring global network state. The authors suggest future extensions such as incorporating energy consumption models, mobility‑prediction based link availability, and support for multiple QoS classes to further enhance the robustness and applicability of CARP in real‑world ad‑hoc deployments.