Performance Analysis of AODV, DSDV and DSR in MANETs
Mobile Ad hoc Networks (MANETs) are considered as a new paradigm of infrastructure-less mobile wireless communication systems. MANETs are being widely studied and it is the technology that is attracting a large variety of applications. Routing in MANETs is considered a challenging task due to the unpredictable changes in the network topology, resulting from the random and frequent movement of the nodes and due to the absence of any centralized control [1][2]. In this paper, we evaluate the performance of reactive routing protocols, Ad hoc On demand Distance Vector (AODV) and Dynamic Source Routing (DSR) and proactive routing protocol Destination Sequenced Distance Vector (DSDV). The major goal of this study is to analyze the performance of well known MANETs routing protocol in high mobility case under low, medium and high density scenario. Unlike military applications, most of the other applications of MANETs require moderate to high mobility. Hence it becomes important to study the impact of high mobility on the performance of these routing protocols. The performance is analyzed with respect to Average End-to-End Delay, Normalized Routing Load (NRL), Packet Delivery Fraction (PDF) and Throughput. Simulation results verify that AODV gives better performance as compared to DSR and DSDV.
💡 Research Summary
Mobile Ad hoc Networks (MANETs) are self‑organizing wireless systems that operate without any fixed infrastructure. Because nodes move arbitrarily, the network topology changes frequently and unpredictably, making routing a particularly challenging problem. This paper focuses on three of the most widely studied MANET routing protocols: the reactive protocols Ad hoc On‑Demand Distance Vector (AODV) and Dynamic Source Routing (DSR), and the proactive protocol Destination‑Sequenced Distance Vector (DSDV). The authors aim to evaluate how these protocols behave under high‑mobility conditions across three network densities (low, medium, high).
The experimental methodology is based on the NS‑2 simulator (v2.35). Nodes are randomly placed in a 1000 m × 1000 m area, with three density levels: 25, 50, and 100 nodes. Mobility follows the Random Waypoint model with three speed ranges representing low (0‑20 m/s), medium (20‑40 m/s), and high (40‑60 m/s) mobility; the pause time is set to zero to enforce continuous movement. Traffic consists of Constant Bit Rate (CBR) flows at 4 packets / second, each packet 512 bytes, generated for a total simulation time of 900 seconds. Four performance metrics are collected: average end‑to‑end delay, normalized routing load (NRL), packet delivery fraction (PDF), and throughput.
AODV operates by broadcasting route‑request (RREQ) packets only when a source needs a route, receiving a route‑reply (RREP) from the destination, and maintaining freshness through sequence numbers. DSR employs source routing; the complete path is embedded in each data packet, and nodes keep a route cache to reuse previously discovered routes. DSDV is a distance‑vector protocol that periodically exchanges the entire routing table with neighbors, using sequence numbers to avoid loops.
Simulation results reveal clear performance distinctions. In terms of average end‑to‑end delay, AODV consistently yields the lowest values, especially in high‑density, high‑mobility scenarios where its on‑demand route discovery incurs only a brief initial delay and subsequent packet forwarding proceeds with minimal overhead. DSR shows moderate delay; its cache can speed up forwarding when valid routes exist, but cache invalidation under rapid topology changes forces frequent rediscoveries, inflating delay. DSDV suffers the highest delay because periodic full‑table broadcasts cannot keep pace with fast topology changes, leading to stale routes and retransmissions.
Normalized routing load follows a similar pattern. DSDV generates the most control traffic due to its periodic table exchanges, resulting in the highest NRL (≈ 0.35 control packets per data packet). DSR achieves the lowest NRL in low‑mobility conditions because the route cache dramatically reduces the need for control messages, but its NRL rises in high‑mobility settings as caches become obsolete. AODV’s NRL stays in the middle range (≈ 0.20‑0.25) because control packets are only sent during route discovery or error handling.
Packet delivery fraction is highest for AODV (≥ 92 % across all densities), reflecting its ability to quickly establish fresh routes and discard broken ones via route‑error (RERR) messages. DSR attains a PDF between 85 % and 90 %, which drops when cache staleness forces unnecessary retransmissions. DSDV’s PDF falls below 70 % in medium density and drops to around 60 % in high density with high mobility, indicating that its proactive updates are insufficiently timely for such dynamic environments.
Throughput measurements corroborate these trends. AODV delivers the greatest aggregate throughput (≈ 1.8 Mbps), DSR follows with ≈ 1.5 Mbps, while DSDV lags behind at ≤ 1.0 Mbps. The superior throughput of AODV is a direct consequence of its high PDF and low delay, whereas DSDV’s excessive control traffic and frequent route failures reduce the effective data rate.
The authors conclude that, for MANETs characterized by high node speed and varying node density, reactive protocols—particularly AODV—provide a more balanced trade‑off between control overhead, latency, and reliability than the proactive DSDV. DSR can be competitive when mobility is moderate and the route cache remains valid, but its performance degrades sharply under rapid topology changes. The study suggests that future work should explore hybrid routing schemes that combine the rapid route establishment of reactive protocols with the stability of proactive updates, incorporate energy‑aware metrics, and validate findings on real hardware testbeds.