An Energy-Aware Multi-Packet Forwarding and Mobility-Resilient Caching Model for IoT-Enabled Named Data Ad Hoc Networks

Introduction

The traditional Internet follows a host-centric communication model in which data are exchanged between network endpoints identified by addresses. This model establishes communication with a particular network location before the requested information can be retrieved. Many current Internet applications, however, are centered on accessing content rather than communicating with a specific host. In such cases, the primary requirement is to obtain the requested content, regardless of the network location from which it is delivered. Applications such as video streaming, social networking, cloud storage, mobile services, and Internet of Things (IoT)-based sensing generate and consume massive amounts of named content across heterogeneous and mobile environments. The conventional TCP/IP architecture has a number of limitations in these situations, such as location dependence, inadequate support for in-network caching, ineffective content retrieval, scalability problems, costs associated with mobility management, and duplicated data transfer.

By replacing content-based naming for host-based addressing, NDN has been proposed as a potential future Internet architecture to get over these limitations. An NDN consumer provides an Interest packet with the name of the desired content instead of sending packets to a destination IP address. The necessary data packet may be returned by the producer or any intermediary node that has a cached matching copy. Because of this, NDN is a perfect match for networks with limited resources, mobile communication, and content delivery. NDN architecture is based on two types of packets, Interest and Data, and three main data structures on routers, the Content Store (CS), Pending Interest Table (PIT) and Forwarding Information Base (FIB). The CS briefly caches the incoming Data packets, the PIT records the unsatisfied Interests and the FIB forwards the Interests towards possible sources of content [1]. The NDN project description also states that if several Interests for the same data are received from downstream interfaces, only the first Interest is transmitted upstream, and returning Data is forwarded to all requesting interfaces and cached in the CS.

These features make NDN suitable for wireless ad hoc networks such as Mobile Ad Hoc Networks (MANETs), Vehicular Ad Hoc Networks (VANETs), IoT networks and wireless sensor networks. In these networks the nodes connect without a fixed infrastructure and each node might be a consumer, a producer or a relay. NDN allows intermediary nodes to cache and return content regardless of the location of the original producer, and therefore it can reduce delay of retrieval and enhance availability of content. Recent research further highlight the advantages of NDN-based wireless ad hoc networks such as in-network caching, mobility support, scalability, security, and location-independent data access [1-3]. However, these benefits are not obtained automatically. The broadcast nature of wireless communications, frequent mobility of nodes, limited battery capacity, and unpredictable connectivity create various performance difficulties.

Packet flooding is a major problem in NDN-based ad hoc networks. In a wired NDN network, Interest forwarding normally takes relatively stable channels using FIB-based forwarding, whereas Data packets come back along the reverse path built in PIT. On the other hand, wireless ad hoc communication is broadcast-based. A node flooding an Interest packet might reach numerous neighbors. Excessive Interest flooding, redundant Data transmission, packet collision, and unnecessary energy consumption may occur if these neighbors re-broadcast the same Interest packets or return duplicate Data packets. The problem is multiplied when the same content is broken down into many Data packets and the user has to send an Interest for each packet. We highlight that the traditional "one Interest for one Data packet" behavior is a main reason for inefficiency of round-trip time and redundant forwarding in wireless NDN setups.

Another major restriction is the operation efficiency of CS and PIT. Each arriving Interest requires a lookup in the CS to verify whether the requested Data is locally available. If no matching Data is available, the node checks the PIT to see if the same Interest is pending. Similarly when a Data packet comes, the PIT must be checked to identify the requested downstream interfaces and the Data may also be inserted into the CS for future use. Therefore, lookup, insertion, update and deletion operations in CS and PIT directly influence the forwarding delay and node processing overhead.

Mobility adds a further aspect to the complexity of ad hoc communication based on NDN. In a mobile ad hoc network, the positions of consumers, producers and intermediate nodes may shift throughout the retrieval of content. A consumer may send an Interest from one location and then relocate before all Data packets are received. Similarly, a producer or intermediate node may move out of transmission range during the time a request is partially satisfied. In this instance, the consumer might receive just some packets of a multi-packet content object, while other packets might be lost, or delayed until the relevant PIT entries expire. The significance of the problem is also stressed in recent NDN mobility literature, where the dynamic movement of producers and consumers impacts on the handoff latency, the efficiency of packet delivery and the performance of data retrieval [4, 5]. We takes into consideration a consumer mobility scenarios when a consumer shifts positions before all data packets arrive, resulting in an incomplete or only partially completed request.

Another important issue with wireless ad hoc networks is energy usage. Nodes in sensor-based systems, MANETs, and IoT networks frequently run on constrained battery power. If every node continuously participates in Interest forwarding, Data forwarding, caching, and retransmission, low-energy nodes may deplete their batteries quickly. This can reduce node lifetime and may also damage network connectivity. Existing energy-aware NDN forwarding schemes, such as OEFS, consider residual energy during forwarding decisions [6]. Other recent works have also continued to investigate smart and energy-aware forwarding for NDN-based environments, confirming that energy-efficient forwarding remains a relevant research direction [7,8]. However, many energy-aware approaches either focus on single-packet forwarding or require additional control messages to exchange energy information, which may itself increase overhead.

Several existing schemes have attempted to address forwarding, caching, energy, or mobility issues in NDN-based wireless networks. Blind Forwarding (BF) uses simple broadcast forwarding but may generate excessive redundant packets. LOMCF improves forwarding and caching in NDN-based MANETs by using location-aware multipath forwarding and caching decisions [3]. REF reduces redundant PIT entries in wireless NDN, but its forwarding process is mainly based on single-packet requests [9]. OEFS considers the residual energy of a node before forwarding an Interest and is intended to reduce unnecessary energy consumption in wireless NDN ad hoc networks [6]. Other studies address different parts of the same problem. Multipath Interest forwarding is considered in [10], cache replacement and cache organization are studied in [11,12], and mobility recovery is addressed in [13,14]. More recent studies also report that cache placement, limited storage, mobility, and resource constraints remain relevant issues in wireless NDN and IoT environments [15,16]. The studies reviewed above generally address forwarding, caching, mobility, or energy management as separate functions. Most of them also retain the conventional one-Interest-one-Data exchange, which requires a separate request for each Data packet.

A closely related paper for our work is E-DRAFT [17]. E-DRAFT uses an accumulative Interest mechanism for NDN-based wireless multimedia sensor networks and employs an array-based B+Tree Content Store to improve Data retrieval and control packet flooding. The present work follows the same general direction of reducing unnecessary packet exchange, but considers a different forwarding problem. It maintains the status of individual Data packets inside an Interest and combines this information with hash-based CS/PIT access, residual-energy checking, delayed Interest processing, and consumer mobility recovery. These functions are intended for IoT-enabled ad hoc networks where packet retrieval may be affected by both node energy and movement. Based on these observations, we develop an energy-aware multi-packet forwarding and cache-assisted mobility model for NDN-based ad hoc networks. A consumer does not generate a separate Interest for every Data packet. Instead, one Interest carries a packet-status map that records the packets that are still required and those that have already been received. An intermediate node checks this map before forwarding the Interest. It can satisfy available packets from its local CS, update the corresponding PIT information, and forward the request only for packets that remain unresolved.

The forwarding process also considers the operating condition of the intermediate node. CS and PIT entries are maintained through hash-based structures to support frequent lookup and update operations. The node checks a freshly received Interest’s remaining energy against a predetermined threshold before transmitting it. After an Interest is approved, it is stored for a brief period of time so that requests coming in from various routes can be pooled before being forwarded onward. When a continuous retrieval is still pending, mobility is taken into consideration. An additional Interest is produced for the packets that are still unresolved rather than for the entire content if the consumer travels farther than the set distance threshold. The main contributions of this paper are summarized as follows:

• We introduce a multi-packet forwarding mechanism for NDN-based ad hoc networks in which a single Interest can represent the retrieval state of several Data packets rather than requesting each packet independently.

• We add packet-level retrieval state into Interest processing through a status map that distinguishes received packets from unresolved packets. Intermediate nodes use this state during PIT updates, cache lookup, and subsequent Interest forwarding.

• We organize CS and PIT entries using hash-based structures to support frequent content lookup and packet-status updates with average-case O(1) access.

• We add residual energy into the Interest-forwarding decision. A node whose available energy is below the configured threshold does not accept a new Interest-forwarding task.

• A short delayed-processing interval is used to combine overlapping Interests received through different paths before they are forwarded further.

• Consumer mobility is handled by monitoring displacement from the request location. When the configured threshold is exceeded, only the packets that remain missing are requested again.

• The model is implemented and evaluated in ndnSIM under ad hoc network conditions using packet-transmission overhead, content retrieval time, CS/PIT lookup, node lifetime, and mobility-related measurements.

The remainder of this paper is organized as follows. Section 2 discusses previous work on NDN forwarding, caching, energy-aware communication, mobility, and CS/PIT organization. Section 3 defines the network model and formulates the problem. Section 4 describes the proposed forwarding model and its packet-processing procedure. Section 5 presents the simulation configuration and evaluation metrics. The results are discussed in Section 6, followed by limitations and future work in Section 7. Section 8 concludes the paper.

Complete Article

The complete article, including all figures, tables, equations and algorithms, is available in the official publication PDF.

View Full PDF

Conclusion

This paper investigated multi-packet content retrieval in IoT-enabled NDN ad hoc networks under energy and mobility constraints. The proposed model combines a packet-status-aware one-Interest-multiple-packets mechanism with hash-based CS/PIT management, delayed Interest merging, residual-energy-based forwarding, and mobility-triggered re-requesting. Under the evaluated scenarios, the proposed model forwarded fewer Interest and Data packets and achieved a lower content retrieval time than BF. It also recorded a longer node lifetime over the evaluated network sizes. For CS/PIT access, the hash-based implementation produced lower measured search times than the compared B-tree and set structures as the number of stored entries increased. The model differs from compared approach architecture by maintaining the retrieval state of individual packets and by incorporating explicit energy-based forwarding and consumer-mobility recovery. These results indicate that packet-level retrieval state can be used to coordinate forwarding, caching, and recovery decisions in resource-constrained NDN ad hoc networks.

Future work will investigate more compact representations of the packet-status map and adaptive selection of the packets-per-Interest value.

References

1. A. Tariq, R. A. Rehman, and B.-S. Kim, “Forwarding strategies in NDN-based wireless networks: A survey, ” IEEE Communications Surveys & Tutorials, vol. 22, no. 1, pp. 68–95, 2020.
2. Y. Yu, R. B. Dilmaghani, S. Calo, M. Y. Sanadidi, and M. Gerla, “Interest propagation in named data MANETs, ” in Proc. International Conference on Computing, Networking and Communications, 2013, pp. 1118–1122.
3. R. A. Rehman and B.-S. Kim, “LOMCF: Forwarding and caching in Named Data Networking based MANETs, ” IEEE Transactions on Vehicular Technology, vol. 66, no. 10, pp. 9350–9364, 2017.
4. W. M. H. Azamuddin, A. H. M. Aman, H. Sallehuddin, K. Abualsaud, and N. Mansor, “Mathematical models for Named Data Networking producer mobility techniques: A review, ” Mathematics, vol. 12, no. 5, Art. no. 649, 2024, doi: 10.3390/math12050649
5. Z. Liu et al., “MsDD: A novel NDN producer mobility support scheme based on multi-satellite data depot, ” PLOS ONE, 2024, doi: 10.1371/journal.pone.0310379
6. R. A. Rehman, S. H. Ahmed, and B.-S. Kim, “OEFS: On-demand energy-based forwarding strategy for named data wireless ad hoc networks, ” IEEE Access, vol. 5, pp. 6075–6086, 2017.
7. N. A. Askar, A. Habbal, H. Hamouda, A. M. Alnajim, and S. Khan, “SEF: A smart and energy-aware forwarding strategy for NDN-based Internet of Healthcare, ” Computers, Materials & Continua, vol. 81, no. 3, pp. 4625–4658, 2024, doi: 10.32604/cmc.2024.058607
8. D. Papenfuss, M. T. R. Khan, et al., “Enhancing energy efficiency in IoT-NDN via parameter optimization, ” Future Internet, vol. 16, no. 2, Art. no. 61, 2024.
9. N. N. Dehaghani, R. Sadeghi, and M. F. Imani, “REF: A novel forwarding strategy in wireless NDN, ” Wireless Personal Communications, vol. 117, no. 3, 2021, doi: 10.1007/s11277-020-07909-8
10. R. A. Rehman, T. D. Hieu, H. M. Bae, S. H. Mah, and B.-S. Kim, “Robust and efficient multipath Interest forwarding for NDN-based MANETs, ” in Proc. IFIP Wireless and Mobile Networking Conference, 2016, pp. 187–192.
11. M. Meddeb, A. Dhraief, A. Belghith, T. Monteil, K. Drira, and H. Mathkour, “Least fresh first cache replacement policy for NDN-based IoT networks, ” Pervasive and Mobile Computing, vol. 52, pp. 60–70, 2019, doi: 10.1016/j.pmcj.2018.12.002
12. A. T. Herouala, B. Ziani, C. A. Kerrache, A. E. K. Tahari, N. Lagraa, and S. Mastorakis, “CaDaCa: A new caching strategy in NDN using data categorization, ” Multimedia Systems, 2022, doi: 10.1007/s00530-022-00904-y
13. M. Meddeb et al., “AFIRM: Adaptive forwarding based link recovery for mobility support in NDN/IoT networks, ” Future Generation Computer Systems, vol. 87, pp. 351–363, 2018.
14. J. H. Cha et al., “A mobility link service for NDN consumer mobility, ” Wireless Communications and Mobile Computing, 2018.
15. A. Khalid, R. A. Rehman, M. Burhan, et al., “Caching strategies in NDN based wireless ad hoc network: A survey, ” Computers, Materials & Continua, vol. 80, no. 1, pp. 1469–1496, 2024, doi: 10.32604/cmc.2024.054448
16. A. A. Alahmad, A. H. M. Aman, F. Qamar, and W. Mardini, “Efficient caching strategies in NDN-enabled IoT networks: Strategies, constraints, and future directions, ” Sensors, vol. 25, no. 16, Art. no. 5203, 2025, doi: 10.3390/s25165203
17. A. Arsalan, M. Burhan, R. A. Rehman, T. Umer, and B.-S. Kim, “E-DRAFT: An efficient data retrieval and forwarding technique for Named Data Network based wireless multimedia sensor networks, ” IEEE Access, vol. 11, pp. 15315–15328, 2023, doi: 10.1109/ACCESS.2023.3244247
18. M. M. Qazi, R. A. Rehman, and B.-S. Kim, “Minimizing content-store data access time using two-tier tree architecture for NDN-based WSNs, ” in Proc. IEEE Wireless Communications and Networking Conference Workshops, 2020, pp. 1–6.
19. S. Gao, H. Zhang, and B. Zhang, “Energy efficient Interest forwarding in NDN based wireless sensor networks, ” Mobile Information Systems, vol. 2016, 2016.
20. A. Aboud, H. Touati, and B. Hnich, “Efficient forwarding strategy in a NDN-based Internet of Things, ” Cluster Computing, vol. 22, no. 4, 2019, doi: 10.1007/s10586-018-2859-7