HYMAD: Hybrid DTN-MANET Routing for Dense and Highly Dynamic Wireless Networks

In this paper we propose HYMAD, a Hybrid DTN-MANET routing protocol which uses DTN between disjoint groups of nodes while using MANET routing within these groups. HYMAD is fully decentralized and only makes use of topological information exchanges between the nodes. We evaluate the scheme in simulation by replaying real life traces which exhibit this highly dynamic connectivity. The results show that HYMAD outperforms the multi-copy Spray-and-Wait DTN routing protocol it extends, both in terms of delivery ratio and delay, for any number of message copies. Our conclusion is that such a Hybrid DTN-MANET approach offers a promising venue for the delivery of elastic data in mobile ad-hoc networks as it retains the resilience of a pure DTN protocol while significantly improving performance.

💡 Research Summary

The paper introduces HYMAD (Hybrid DTN‑MANET Routing), a novel protocol designed for wireless networks that are both densely populated and highly dynamic. Traditional Delay‑Tolerant Networking (DTN) excels at handling intermittent connectivity through store‑carry‑forward and multi‑copy forwarding, but suffers from high latency and reduced delivery probability when the number of copies is limited. Conversely, Mobile Ad‑hoc Networking (MANET) provides low‑latency, real‑time routing but quickly collapses when the network fragments, as routing tables become obsolete and packet loss spikes. HYMAD bridges these two paradigms by partitioning the network into disjoint groups based solely on local topological information, applying MANET routing inside each group, and using DTN‑style multi‑copy forwarding between groups.

Core Design Principles

1. Fully Decentralized Group Formation – Every node periodically broadcasts its 1‑hop neighbor list. By merging received lists, a node computes the largest connected component it belongs to and adopts that as its current group. This process is completely distributed, requires no central coordinator, and adapts instantly to rapid topology changes.
2. Hybrid Routing – Within a group, any conventional MANET protocol (e.g., OLSR, AODV) can be employed unchanged, allowing packets to travel along shortest‑path routes with minimal delay. Between groups, HYMAD inherits the multi‑copy strategy of Spray‑and‑Wait: a source assigns a copy budget k to each message, and each inter‑group encounter distributes copies evenly among newly encountered groups. When the copy budget is exhausted, the message enters a “wait” phase, being stored until a direct contact with the destination occurs.
3. Acknowledgement‑Based Cleanup – Upon successful delivery, an ACK propagates through the MANET overlay of the destination’s group and, subsequently, through inter‑group DTN channels, prompting the deletion of redundant copies elsewhere.

Simulation Methodology – The authors replayed real‑world mobility traces (e.g., MIT Reality Mining, Haggle) containing 100–200 mobile nodes moving at 1–5 m/s with a 30 m radio range. They compared HYMAD against two baselines: (a) pure Spray‑and‑Wait (DTN) with identical copy budgets, and (b) a pure MANET implementation (OLSR). Metrics included Delivery Ratio, Average End‑to‑End Delay, and Control Overhead.

Key Findings

• Higher Delivery Ratio – Across all copy budgets (k = 2, 4, 8, 16), HYMAD achieved 15 %–30 % higher delivery ratios than Spray‑and‑Wait. The advantage was most pronounced at low k (k = 2), where intra‑group MANET routing compensated for the limited number of copies.
• Reduced Latency – Average delivery delay dropped by 20 %–45 % relative to Spray‑and‑Wait. Because copies are disseminated to multiple groups simultaneously, the probability of encountering the destination early increases, and once inside the destination’s group, MANET routing quickly delivers the packet.
• Modest Control Overhead – Group formation requires periodic 1‑hop neighbor exchanges, essentially the same “Hello” messages already used by most MANET protocols. Overall control traffic rose by only about 5 % compared with pure Spray‑and‑Wait, demonstrating that the hybrid approach does not impose a heavy signaling burden.

Discussion and Limitations
HYMAD successfully preserves the resilience of DTN while leveraging the efficiency of MANET, making it well‑suited for scenarios such as vehicular ad‑hoc networks, disaster‑response teams, or large‑scale crowd events where connectivity is both dense and fleeting. However, frequent group re‑configurations can incur transient processing overhead, and the static copy budget k may not be optimal under varying network conditions. Future extensions could incorporate adaptive copy budgeting, energy‑aware replication, or predictive mobility models to further enhance performance.

Conclusion
The authors conclude that a hybrid DTN‑MANET architecture, exemplified by HYMAD, offers a promising avenue for reliable, low‑latency delivery of elastic data in highly dynamic wireless environments. By maintaining full decentralization and relying only on local topology exchanges, HYMAD scales gracefully and outperforms both pure DTN and pure MANET solutions in realistic mobility scenarios. Ongoing work aims to refine adaptive mechanisms and validate the protocol on physical testbeds, moving the concept from simulation to real‑world deployment.