An Optimized Probabilistic Delay Tolerant Network (DTN) Routing Protocol Based on Scheduling Mechanism for Internet of Things (IoT)

Yuxin Mao, Chenqian Zhou, Yun Ling, Jaime Lloret2019 · DOI

TL;DR

This paper introduces Scheduling-PROPHET, a novel routing protocol for Delay Tolerant Networks (DTNs) within the Internet of Things (IoT). By enhancing the traditional PROPHET protocol with intelligent message scheduling and path selection mechanisms, it significantly improves message delivery rates and reduces delays, especially in challenging, intermittently connected environments.

What Problem Does This Paper Address?

The Internet of Things (IoT) is rapidly expanding, connecting a vast array of devices, many of which operate in environments characterized by unreliable wireless connectivity, dynamic mobility, and frequent link interruptions. These conditions make traditional networking protocols ineffective, necessitating the use of Delay Tolerant Networks (DTNs). DTNs rely on a "store-carry-forward" mechanism, where nodes temporarily store and forward messages. However, existing DTN routing protocols, even probabilistic ones like PROPHET, face challenges such as message loss due to buffer congestion, dropped messages from intermediate nodes, and inefficient path selection. This paper addresses the critical need for efficient and reliable data transmission in IoT applications operating within these challenging DTN environments.

Methods

The core contribution of this paper is the Scheduling-PROPHET protocol, an extension of the existing PROPHET routing protocol. The enhancements focus on two key areas:

  • Delivery Predictability Calculation: Building upon PROPHET's use of encounter history, Scheduling-PROPHET refines the calculation of delivery predictability. It uses an incremental averaging method and considers the cumulative predictability of paths, not just direct node-to-node encounters.
  • Scheduling Mechanisms: Two novel scheduling mechanisms are introduced:
  • Message Management Mechanism: This mechanism prioritizes messages based on their delivery predictability. Lower priority messages are managed to avoid buffer overflow and reduce unnecessary replication, thereby optimizing resource utilization.
  • Message-Transfer Mechanism: This mechanism optimizes path selection. When a node encounters another, it considers multiple factors:
  • Highest Delivery Predictability: The primary criterion is selecting the path with the highest cumulative predictability to the destination.
  • Minimum Hop Count: If multiple paths share the highest predictability, the one with the fewest hops is chosen.
  • Least Traffic Load: If paths are still tied after considering hop count, the path with the least traffic load on the next hop is selected.

These mechanisms aim to improve transmission efficiency, reduce message loss, and minimize delivery delay.

Datasets

Not discussed in the paper. The paper focuses on simulation-based evaluation.

Key Insights and Ideas

  • Predictability is Key, but Context Matters: While encounter history (predictability) is crucial for DTN routing, simply relying on direct encounters isn't enough. Considering the cumulative predictability of a path and incorporating secondary factors like hop count and traffic load leads to more robust and efficient routing.
  • Intelligent Buffering Reduces Waste: Prioritizing messages within a node's buffer based on their likelihood of delivery prevents valuable buffer space from being occupied by low-priority or unlikely-to-be-delivered messages, thus improving overall network efficiency.
  • Multi-Criteria Path Selection is Superior: A single-criterion path selection (e.g., only predictability) can be suboptimal. Combining predictability with hop count and traffic load provides a more holistic approach to finding the best route in dynamic DTN environments.
  • IoT Demands DTN Solutions: The paper strongly reinforces that the inherent characteristics of many IoT deployments (mobility, intermittent connectivity) necessitate DTN approaches, and protocols like Scheduling-PROPHET are vital for their success.

Key Results

The paper presents simulation results using the ONE simulator, comparing Scheduling-PROPHET against Epidemic and PROPHET. Key findings include:

  • Improved Delivery Rate: Scheduling-PROPHET consistently achieved higher delivery rates across various scenarios, especially under conditions of high node density and longer transmission ranges.
  • Reduced Delivery Delay: The protocol demonstrated significantly lower average delivery delays, particularly in high-density networks.
  • Lower Overhead: The message management mechanism helped reduce network overhead by avoiding unnecessary message replication and prioritizing transmissions.
  • Fewer Dropped Packets: Scheduling-PROPHET exhibited a lower number of dropped packets, attributed to its efficient buffer management and intelligent path selection.
  • Performance Across Scenarios: The protocol showed superior performance across different transmission ranges, node densities, and buffer sizes, indicating its adaptability.

Limitations and Future Work

  • Buffer Size Impact: The authors note that while Scheduling-PROPHET performs better with increasing buffer sizes, a larger buffer is generally beneficial for achieving optimal performance in real-life IoT applications.
  • Bluetooth Dependency: The simulations were conducted with Bluetooth as the transmission interface. While this is common in some IoT scenarios, future work could explore performance with other communication technologies.
  • Further Optimization: The paper suggests that extending the existing PROPHET protocol with their proposed mechanisms is a promising avenue. Future work could explore further optimizations and real-world deployments to validate these findings.

Summary updated

BibTeX
@misc{mao2019optimized,
  author = {Yuxin Mao and Chenqian Zhou and Yun Ling and Jaime Lloret},
  title = {{An Optimized Probabilistic Delay Tolerant Network (DTN) Routing Protocol Based on Scheduling Mechanism for Internet of Things (IoT)}},
  year = {2019},
  doi = {10.3390/s19020243},
}
Figure