Elsevier

Ad Hoc Networks

Volume 94, November 2019, 101936
Ad Hoc Networks

Virtual mesh networking for achieving multi-hop D2D communications in 5G networks

https://doi.org/10.1016/j.adhoc.2019.101936Get rights and content

Abstract

In 5G networks, when D2D communications or the quality of a cellular link fails to support data transmission of a terminal, it is indispensable to exploit data transmission opportunities via multi-hop D2D communications. However, to achieve multi-hop D2D communications, two problems need to be solved: 1) A routing and packet forwarding protocol is required to enable terminals to forward data packets across multi-hop D2D or cellular links; 2) This protocol must conform to the existing specifications of D2D communications. To this end, a framework of virtual mesh networking is developed in this paper to support multi-hop D2D communications in 5G networks. More specifically, based on the 5G network architecture, routing and packet forwarding are split into two separate mechanisms in the control plane and the user plane, respectively. In the control plane, base stations collect link and topology information of terminals and then form a virtual mesh network among terminals accordingly. The routing path for end-to-end communications is then determined by a virtual mesh routing algorithm, which also takes into account terminal mobility and link failure. In the user plane, a packet forwarding mechanism is designed to deliver data packets following the routing path as determined by the routing algorithm in the control plane. This mechanism extends but also conforms to the existing specifications of D2D communications. The virtual mesh networking protocol and its underlying algorithms are evaluated via simulations. Performance results demonstrate the effectiveness of multi-hop D2D communications via virtual mesh networking.

Introduction

D2D communications are promising for high speed direct communications when two terminals are in the proximity of each other [1]. It has been specified in 3GPP standard in Release 12 and further strengthened in Release 13–15. Two types of communication scenarios are specified in Release 13 [2]: 1) single-hop direct data transmission between two terminals, as shown in scenario 1 in Fig. 1; 2) two-hop data transmission between a terminal and the core network using another terminal as a relay, as shown in scenario 2 in Fig. 1. When two terminals are not in the proximity of each other, they have to communicate through the core network via cellular links. However, such links may not be available in two typical scenarios: 1) small cell base stations lack wireless resources to support massive access in large gatherings; 2) communication infrastructure crashes at the occurrence of natural disasters. In these situations, multi-hop D2D communications are indispensable to exploit data transmission opportunities. Such opportunities are shown in scenarios 3–5 in Fig. 1, where a two-hop link can be further extended to a multi-hop link.

Multi-hop D2D communications play an important role in improving end-to-end connectivity between two terminals. This is particularly critical for providing communication opportunities in public safety applications, as studied in [3] and analyzed in [4]. Moreover, it is crucial when mmWave is used for multimedia applications [5] in 5G access networks, since mmWave suffers from high attenuation during transmission and can be blocked by the surrounding objects. Moreover, multi-hop D2D communications also expand the network coverage significantly, which is essential to support massive devices in Internet of Things (IoT) applications [6]. Although the UE-to-network relay technology for IoT and wearables has been enhanced in [7], multi-hop D2D communications have not been specified in the latest 3GPP standards so far. Therefore, it is critical to study multi-hop D2D communications in 5G networks.

Two issues need to be addressed in multi-hop D2D communications. First, if a terminal generates a packet or receives a packet not for itself, it needs to send the packet to an appropriate next-hop terminal. Second, when receiving a packet, a terminal needs to determine whether the packet is for itself, and forwards the packet once it is not. Consequently, a routing and forwarding mechanism needs to be designed for multi-hop D2D communications.

Although there exists a routing and packet forwarding mechanism in current D2D standards for uplink data transmission (i.e., scenario 2 in Fig. 1), it cannot be extended to support multi-hop D2D communications for the following reason. In this mechanism, evolved packet switched system (EPS) bearers [8] to the core network are first established at a relay, and a traffic flow template (TFT) [2] is constructed if the relay sets up a D2D link with a distant terminal. When the relay receives a packet from the distant terminal, it sends the packet to the core network through the EPS bearer based on the TFT. Since the EPS bearer is a tunnel from the relay to the core network, the packet is not visible at the base station until it reaches the core network. As a result, this mechanism cannot be leveraged to support routing and packet forwarding in multi-hop D2D communications.

Multi-hop communications have been extensively studied in ad hoc and mesh networks. So far, many routing protocols are available to support multi-hop communications, and most of them can be classified into distributed routing protocols or hybrid routing protocols. In a distributed routing protocol, control messages are distributively disseminated in a network to find the route from a source to a destination. There are two widely used distributed routing protocols: 1) optimized link state routing (OLSR) [9], where the aforementioned procedure is performed proactively; 2) ad hoc on-demand distance vector (AODV) [10], where the routing procedure is performed reactively. In a hybrid routing protocol, each node proactively maintains a routing path to a root node. If a source node cannot find an existing route for a packet, it sends the packet to the root node and triggers a route setup procedure reactively at the same time. As a result, a route is set up and packets can be transmitted via the newly set up route. A typical example of hybrid protocols is hybrid wireless mesh protocol (HWMP), which is specified in IEEE 802.11s [11].

The above two types of routing protocols are not suited for multi-hop D2D communications for two reasons. First, their route setup procedures do not conform to the standardized D2D specifications. In both types of routing protocols, an intermediate terminal broadcasts route setup request messages after it receives one from its neighbors. However, in the current D2D standards, a terminal needs to be triggered by an application and authorized by the ProSe Function Application Server in the core network before conducting broadcasting. As a result, lots of modifications to the current standards are necessary to adopt the existing mesh or ad hoc routing protocols in multi-hop D2D communications. Second, the route setup procedures are not efficient for multi-hop D2D communications. Signaling messages from a source need to go through multiple hops before reaching a destination, resulting in large latency in route setup and route failure recovery. However, such hop-by-hop signaling procedures are not necessary in 5G networks, since signaling can be conducted via a single-hop cellular link, following a path different from that of multi-hop data communications.

Besides the above routing protocols, multipath routing can be another option for routing mechanisms of mesh networks. Typical examples include disjoint multipath routing [12] and braided multipath routing [13]. Multipath routing does not only facilitate load balancing but also improves transmission reliability. However, since transmitting packets via a large number of hops leads to large delay, the number of hops of a routing path in multi-hop D2D communications is actually limited. Consequently, there exist few totally disjoint or partially disjoint routing paths between a given source-destination pair, which makes the benefits of multipath routing marginal. Moreover, to make a multipath routing mechanism effective, associated mechanisms need to be designed [14], e.g., dynamically allocating traffic to different routing paths, reordering packet received from multiple routing paths at the destination, or integrating multipath routing with packet aggregation [15] to reduce overhead. These mechanisms increase the complexity of multi-hop D2D communications and cause high terminal energy consumption when they are executed in terminals. Thus, multi-path routing is not taken into consideration in this paper. However, with the densification of small-cell base stations and the growing interest in integrated access and backhaul (IAB) networks, it is worth future study to consider multipath routing in IAB networks to improve network reliability and perform load balancing dynamically.

Related research work is carried out in [5], [16], [17], [18], [19] to study multi-hop D2D communications. Such work is focused on designing routing algorithms that optimize routes under various constraints such as QoS [5], interference [16], and scheduling [17], [18], [19]. However, how to design routing mechanisms that conform to the standard mechanisms of D2D communications is not considered. Several routing mechanisms surveyed in [20] only consider multi-hop D2D communications as a type of ad hoc networks, and thus they are inapplicable to multi-hop D2D communications. Consequently, a new routing and packet forwarding mechanism that comply with the existing D2D standards needs to be developed to support multi-hop D2D communications.

In this paper, a routing and packet forwarding mechanism is developed to achieve multi-hop D2D communications via standardized one-hop cellular or D2D communication procedures, and it is critical for both low-frequency and mmWave 5G networks. The underlying routing algorithm selects a routing path that may contain a mixture of D2D links and cellular links, but it neglects the interference between cellular and D2D communications. Such a design is valid for two reasons. First, D2D communications can be carried out in the spectrum orthogonal to that of cellular communications, which is called the overlay mode in [21]. Second, resource allocation can be considered separately from routing in a different algorithm, as shown in [22]. To improve the efficiency of the routing mechanism, a routing algorithm that jointly considers routing, resource allocation, and interference needs to be developed, which is an interesting topic subject to future research.

To develop the routing and packet forwarding mechanism, key features of the hierarchical network architecture of 5G networks are leveraged. First, communications between a small-cell base station and terminals or between a macro-cell base station and the small-cell base station can be accomplished within one hop. Thus, neighbor and link information of terminals can be collected and managed by the small-cell base station, and such information is then reported to the macro-cell base station. Based on such information, a virtual mesh network of terminals can be formed in small-cell base stations to accomplish multi-hop D2D communications within a cell, and virtual mesh networks of multiple cells are integrated in the macro cell base station to achieve inter-cell multi-hop communications. Second, signaling messages between terminals can be exchanged via base stations, so they do not have to follow the same path as that of data communications between terminals. In this way, the control plane and the data plane of multi-hop D2D communications are separated to improve protocol efficiency and reliability.

Based on the above two features, the new routing and packet forwarding protocol is designed as follows:

  • A virtual mesh network involving multi-hop D2D communications is formed and managed by macro-cell or small-cell base stations.

  • A virtual mesh routing algorithm is developed based on the virtual mesh network to determine an routing path that may consist of a mixture of cellular links and D2D links.

  • A packet forwarding mechanism conforming to the standard single-hop cellular and D2D communication procedures is designed for a terminal to forward packets according to the routing path determined by the virtual mesh routing algorithm.

Moreover, the challenging issues related to terminal mobility and link failure are also addressed in the new protocol.

The virtual mesh routing and forwarding mechanism developed in this paper is evaluated via extensive simulations. Performance results demonstrate its effectiveness in supporting multi-hop D2D communications. Moreover, it is adaptive to terminal mobility and is also resilient to link failure. Besides, the framework of virtual mesh networking conforms to the existing 3GPP D2D standards.

The rest of the paper is organized as follows. The system model is described in Section 2. The virtual mesh network formation and maintenance procedure is presented in Section 3. A routing algorithm is designed in Section 4, while the corresponding routing mechanism and route failure recovery procedures are illustrated in Section 5. The virtual mesh packet forwarding mechanism is also studied in Section 5. The performance of virtual mesh networking for multi-hop D2D communications is evaluated and analyzed in Section 6. The paper is concluded in Section 7.

Section snippets

System model

In this paper, we consider a multi-hop D2D communication scenario in 5G networks, where there exists a macro-cell base station, several small-cell base stations, and a number of D2D enabled terminals. As shown in Fig. 2(a), macro-cell base station 1 exchanges information with small-cell base stations 1 and 2 through the core network, while small-cell base station 1 communicates with a terminal through a one-hop cellular link. Besides the cellular link, a terminal can also set up a D2D link with

Virtual mesh network formation

In 5G networks, signaling messages are exchanged between a terminal and a base station via a one-hop cellular link. Thus, link and topology information can be quickly collected from the network. Based on such information, a small-cell base station forms and maintains a virtual mesh network for the small cell. It then reports its virtual mesh network topology to a macro-cell base station that integrates the topology from multiple small cells and forms a virtual mesh network for the macro cell.

Virtual mesh routing algorithm

The virtual mesh network G contains all information from the network, so the routing path between any two terminals can be calculated with a routing algorithm. However, two challenges exist in the algorithm design, which makes the existing routing algorithms not effective. First, the mobility of terminals makes a D2D link vulnerable to breakage. Second, an end-to-end routing path may consist of two types of links(i.e., cellular links and D2D links), each with different characteristics. To

Virtual mesh routing protocol

When the source terminal S generates an IP packet but cannot find an existing routing path for the destination terminal D, it sends a route request message to its associated base station, e.g., small-cell base station 1. The message includes the IMSI of terminal S and the IP address of D. Small-cell base station 1 then inquires the PDN Gateway (P-GW) in the core network about the IMSI of D. Since the mapping between the IP address and the IMSI of a terminal is stored in the P-GW, small-cell

Performance evaluation

In this section, the virtual mesh networking protocol and its underlying algorithms are evaluated via simulations. First, the overhead of creating and maintaining virtual mesh networks (i.e., the ratio between the number of signaling messages for creating and maintaining virtual mesh networks and the total number of signaling messages) is analyzed, and the performance of the protocol with respect to different types of link metrics and routing metrics is studied. Second, impact of parameters on

Conclusion

In this paper, a virtual mesh networking mechanism was developed for multi-hop D2D communications to exploit end-to-end data transmission opportunities in 5G networks. It enables a D2D terminal with the capability of forwarding data packets and also conforms to the standard D2D communication specifications. The routing algorithm and key parameters of the virtual mesh routing mechanism were validated via extensive simulations. Performance results showed that the virtual mesh routing mechanism

Conflict of interest

All authors have participated in (a) conception and design, or analysis and interpretation of the data; (b) drafting the article or revising it critically for important intellectual content; and (c) approval of the final version.

This manuscript has not been submitted to, nor is under review at, another journal or other publishing venue.

The authors have no affiliation with any organization with a direct orindirect financial interest in the subject matter discussed in the manuscript.

Acknowledgment

The research work is sponsored by National Natural Science Foundation of China (No. 61771312) and Huawei Technologies, Inc. The authors would like to thank their generous support.

Cheng Huang received the B.S. degree in School of Electronic Information and Communication from Huazhong University of Science and Technology (HUST), Wuhan, China, in 2015. She is currently a Ph.D. student with the Wireless Networking and Artificial Intelligence Laboratory at Shanghai Jiao Tong University. Her current research interests include wireless mesh networking, D2D communications, ultra-dense networks, and 5G networks.

References (33)

  • 3GPP, User equipment (UE) transmission and reception; part 2: range 2 standalone, 2018, (TR 38.101-2,...
  • A. Osseiran

    Scenarios for 5G mobile and wireless communications: the vision of the metis project

    IEEE Commun. Mag.

    (2014)
  • 3GPP, Proximity-based services (ProSe), stage 2, 2016, (TS 23.303, Release 14,...
  • H. Nishiyama et al.

    Relay-by-smartphone : realizing multihop device-to-device communications

    IEEE Commun. Mag.

    (2014)
  • A. Al-Hourani et al.

    Stochastic geometry study on device-to-device communication as a disaster relief solution

    IEEE Trans. Veh. Technol.

    (2016)
  • N. Eshraghi et al.

    Millimeter-wave device-to-device multi-hop routing for multimedia applications

    IEEE International Conference on Communications (ICC)), Kuala Lumpur, Malaysia

    (2016)
  • Qualcomm, Providing the connectivity fabric for everything,...
  • 3GPP, Study on further enhancements to LTE device to device (D2D), user equipment (UE) to network relays for internet...
  • 3GPP, General packet radio service (GPRS) enhancements for evolved universal terrestrial radio access network (E-UTRAN)...
  • T. Clausen, P. Jacquet, Optimized link state routing protocol (OLSR), 2003, (RFC...
  • C. Perkins, E. Belding-Royer, S. Das, Ad hoc on-demand distance vector (AODV) routing, 2003, (RFC...
  • IEEE standard association, Part 11: wireless lan medium access control(MAC) and physical layer (PHY) specifications...
  • S.-J. Lee et al.

    Split multipath routing with maximally disjoint paths in ad hoc networks

    IEEE International Conference on Communications (ICC), Helsinki, Finland

    (2001)
  • D. Ganesan et al.

    Highly-resilient, energy-efficient multipath routing in wireless sensor networks

    ACM SIGCOMM Computer Communication Review, New York, USA

    (2001)
  • S. Mueller et al.

    Multipath routing in mobile ad hoc networks: issues and challenges

    Performance Tools and Applications to Networked Systems (M. Calzarossa and E. Gelenbe eds.), vol. 2965 of Lecture Notes in Computer Science, Springer Berlin Heidelberg

    (2003)
  • G.D. Stasi et al.

    Combining multi-path forwarding and packet aggregation for improved network performance in wireless mesh networks

    Comput. Networks

    (2014)
  • Cited by (15)

    View all citing articles on Scopus

    Cheng Huang received the B.S. degree in School of Electronic Information and Communication from Huazhong University of Science and Technology (HUST), Wuhan, China, in 2015. She is currently a Ph.D. student with the Wireless Networking and Artificial Intelligence Laboratory at Shanghai Jiao Tong University. Her current research interests include wireless mesh networking, D2D communications, ultra-dense networks, and 5G networks.

    Bangzhao Zhai received the B.S. degree in 2015 from Southeast University (SEU), Nanjing, China. He is currently working toward the Ph.D. degree with the Wireless Networking and Artificial Intelligence Laboratory, Shanghai Jiao Tong University (SJTU), Shanghai, China. His research interests include wireless mesh networking, millimeter-wave (mmWave) communications, multiple-input-multiple-output (MIMO) techniques, and visible light communications.

    Dr. Aimin Tang received the B.S. degree in Electrical and Computer Engineering from Shanghai Jiao Tong University (SJTU), Shanghai, China, in 2013. He received his Ph.D. degree from Shanghai Jiao Tong University in March 2018. He is currently a research assistant professor with UM-SJTU Joint Institute, Shanghai Jiao Tong University. His current research interests include full duplex communications, rateless coding, D2D communications, and software defined wireless networks.

    Dr. Xudong Wang is currently a professor with the UM-SJTU Joint Institute, Shanghai Jiao Tong University. He is the director of the Wireless Networking and Artificial Intelligence Lab. He is also an affiliate faculty member with the Electrical Engineering Department at the University of Washington. Since he received the Ph.D. degree in Electrical and Computer Engineering from Georgia Institute of Technology in August 2003, he has been working as a senior research engineer, senior network architect, and R&D manager in several companies. He has been actively involved in R&D, technology transfer, and commercialization of various wireless networking technologies. His research interests include wireless communication networks, smart connected systems, and machine learning. He holds several patents on wireless networking technologies and most of his inventions have been successfully transferred to products. Dr. Wang is an editor for IEEE Transactions on Mobile Computing, IEEE Transactions on Vehicular Technology, Elsevier Ad Hoc Networks. He was the demo co-chair of the ACM International Symposium on Mobile Ad Hoc Networking and Computing (ACM MOBIHOC 2006), a technical program co-chair of Wireless Internet Conference (WICON) 2007, and a general co-chair of WICON 2008. He has been a technical committee member of many international conferences. Dr. Wang is an IEEE Fellow.

    1

    Fellow IEEE.

    View full text