ReviewDevice-to-device content caching techniques in 5G: A taxonomy, solutions, and challenges
Introduction
In the last few years, there has been an enormous increase in wireless data traffic, which is expected to grow many folds in the years to come. The major reason for this data exploration is the increase in the number of smart connected devices and data-intensive applications, which in turn increases the amount of data being generated.
The amount of data is generated from social media, real-time data processing units, and other smart devices, which is increasing at a rapid pace. Keeping in view of these challenges, major telecom operators are facing many challenges with respect to stringent quality of service (QoS) provisions to the end-users. So, the existing cellular networks need to enhance the bandwidth for providing a high data rate to the end-users. This is the reason that the next generation of cellular networks will be based on 5G infrastructure to cater to the needs of the end-users requirements [1], [2]. There is an exponential increase in the usage of various smart applications such as intelligent transportation systems, smart e-healthcare, and smart education with the rapid development of device-to-device (D2D) and machine-to-machine communication [3], [4]. Fig. 1 [5] shows an increase in the number of users with an increase in smart devices usage. The major challenges in 5G are high bandwidth, low latency, high data rates, Device-to-Device (D2D) communication, and an increase in the number of connected devices. D2D communication enables the users to connect directly with each other without having any connectivity with backhaul [6]. D2D treats the user equipments (UE’s) as data hubs for content sharing. To reduce the pressure on the core network, caching at the user’s device for content sharing via D2D communication becomes inevitable.
Caching the popular contents and services at the edge devices brings the content closer to the proximity of the end-user. As the user’s device has a large number of storage capacities, so by using the D2D communication link, these devices can share the popular content with their peers [7]. Using D2D technology, a direct communication link between two User Equipments without the core network can be established. A challenging technique in deploying a D2D communication network is to discover new devices. The device discovery can be performed in either distributed or centralized manner [8], [9], [10]. In contrast to D2D, the traditional communication link must be established using the core network irrespective of whether the two mobile users are in proximity to establish a D2D communication link or not. Both content popularity and the volume of data can be cached, which makes a significant impact on the cached content on the user’s device. It may be possible that at multiple locations same contents can also be cached. To remove the redundancy among the contents caching and sharing, numerous techniques have been discussed in the literature. For example, Nguyen et al. [11] presented a novel chunk based Progressive Popularity-Aware Caching Scheme (PPCS) to improve content availability and to eliminate the cache redundancy issue of Information-Centric Networking. PPCS first cache initial chunks of the content at the edge node and then progressively continues caching subsequent chunks at upstream content nodes along the delivery path, according to the content popularity and each content node position. Psaras et al. [12] proposed a ProbCache scheme to cache the content probabilistically to manage cache resources by reducing the cache redundancy. Zhao et al. [13] proposed a content caching scheme for Vehicular Ad hoc Networks (VANET) using Information-centric networking (ICN). The main objective is to reduce cache redundancy and to lower caching replacement overhead. Moreover, a popularity prediction-based cooperative cache replacement mechanism is proposed, which predicts and ranks popular content during a particular time interval [14], [15]. Naz et al. [16] proposed a coordination based scheme to share the network control information among nodes to take cache decisions. It improves caching with contents diversity and significantly reduces cache redundancy with a trade-off of control traffic overhead. The explicit cache coordination scheme assumes that every router has prior knowledge of network topology, cache’s state, and access frequency. These schemes reduce the cache redundancy but introduce high computational complexity. Rath et al. [17] proposed a scheme that reduces cache redundancy using an off-path central router, which improves the network performance in terms of content duplication and transmission delay. Recently, there has been an increase in the on-the-top (OTT) content. OTT is defined as services that deliver audio and video content over the Internet like Netflix. These services have increased the number of applications
usage. For example, Netflix has increased the number of subscriptions on various platforms, as depicted in Fig. 2(a) [18]. Since 2018, there has been an increase in the number of social media users, as shown in Fig. 2(b) [19]. The need for rapid data rich multimedia has put enormous pressure on the cellular networks. In the previous generations of cellular technologies, the contents were stored at the core network, and all the content access requests were made at the core network. This created a bottleneck situation during peak hours as a result of which latency increased many folds. To address this problem, caching the popular content near the UE’s is proposed in the literature, which results in reducing the delay due to the data transmission among different D2D devices. Fig. 3 shows the comparison between the various generations of cellular technologies as 1G, 2G, 3G, 4G, and 5G. The various testbeds of 5G are yet to be launched in India and other countries. Fig. 3 depicts the evolution of various cellular technologies over the past many years. The transition from 4G to 5G means a shift in the design paradigm from a single discipline system to a multi-discipline system. As illustrated in Fig. 4, 5G supports a wide range of applications, including V2X, where X denotes (V2V (vehicle) or V2I (Infrastructure)), IoT, Smart grid, etc. Massive MIMO will be incorporated at the base stations to reduce latency and further to enhance the data rate. The mm-wave links are exploited for backhauling small cell base stations. The content caching is assisted using Device to Device (D2D) communication. Moreover, by reducing the size of cells, the energy efficiency (EE) of the network can be improved by bringing the network closer to User Equipments (UEs). The key components of 5G are Machine type communication (MTC), Self-Organizing Networks (SON), mm-wave, Backhaul, Energy efficiency (EE), Small cell deployment, Spectrum sharing and Heterogenous networks (HetNet). Caching the popular content near the UE’s results in reducing the delay caused in the transmission of data. The key enabler of 5G is D2D communication, which allows mobile users to access and transfer the cached content through direct wireless connections instead of going through the backhaul network. There are various wireless technologies like Wi-Fi and Bluetooth which enable short-range communications and can help in enabling D2D communication. Fig. 5 illustrates the existing wireless communication technologies used in cellular networks. These technologies differ from each other with respect to the distance between two devices and data rates. For example, Bluetooth 4 supports a 25 Mbps data rate and provides a coverage of 100 m. Moreover, Wi-Fi Direct has a data rate of 250 Mbps with a range of 200 m, and LTE supports a data rate of 14 Mbps having a range of up to 500 m [20]. There are some other short-range wireless communication technologies like Zigbee, SigFox, and LoRaWan (Long Range Wide Area Network), which are best suited for 5G and require multiple devices to be connected at distributed locations. It has been observed that the voluminous amount of data is being generated and transmitted over voice and video calls from the past many years. Thus, there is a need for cost-effective and energy-efficient communication medium. The advantage of using SigFox at the transceiver remains inactive and activates only when data is to be transmitted to save energy. Zigbee has the same bandwidth as Wi-Fi networks but is preferred as it provides more security by limiting the amount of data exchanged and consumes less power. Zigbee is mostly used for industrial applications. LoRaWan is used for connecting the battery-operated devices to the Internet. It provides low power networking and is being preferred for IoT devices [21]. The main focus is to provide mobility, localization, two-way communication, and end-to-end security. Fig. 6 shows the maximum data rates and their coverage area. Hence, content popularity and the volume of data are considered and evaluated in order to identify what contents are to be cached, so as to utilize the finite storage capacities of the UE’s efficiently. D2D communication technology not only ensures the improvement of spectral reuse but also provides low latency, optimal power consumption, fairness, and improved throughput. Therefore, D2D communication technology becomes a key enabler in the upcoming 5G of the cellular networks as it ensures low latency among the end-users.
There are many survey articles in the literature that explore content caching and data sharing among smart devices using D2D technology.
For example, Wang et al. [22] explored caching, computation and communication resources in mobile edge networks. Mao et al. [23] explored the communication perspective of edge computing by considering the latency and energy of the network. Zhang et al. [24] provided a comparative analysis of various frameworks of computational offloading in mobile cloud computing (MCC). Anugraha et al. [25] explored various trust management mechanisms in Mobile Ad Hoc Networks (MANETs) to enhance the performance by managing various attacks among different nodes. Table 1 provides a relative comparison of various existing surveys in the literature.
Currently, a lot of research proposals exist on D2D caching in 5G, which demand high data rates and low latency, so existing proposals face various challenges such as — mobility, privacy and security, resource allocation, and energy-efficiency. These aspects are not exploited to their full potential in the existing solutions. Motivated by these, in this review paper, we analyze and classify different constrains and research challenges in the existing literature on content caching techniques for D2D communication in 5G. The main focus of this research article is on the following issues.
- i.
To design a detailed taxonomy to cover all the aspects of Caching in D2D.
- ii.
To provide a comparison from the existing work on the basis of different parameters of Caching in D2D.
- iii.
To highlight the various research issues and challenges in the existing proposals in the literature.
There are various questions that are required for a detailed review of the literature. Table 2 enumerates these sets of research questions specific to the review article.
The rest of the paper is organized as follows. Section 2 discusses the existing wireless technologies. Section 3 illustrates the systematic planning involved before the review. In Section 4, a detailed taxonomy of D2D content caching in 5G is illustrated by comparing the existing literature on the basis of communication and connectivity, storage, QoS/QoE, and trust management. Section 5 presents the research challenges in D2D content caching techniques in 5G, and finally Section 6 concludes the article. Moreover, Table 3 provides the list of acronyms and their meaning used in the paper.
Section snippets
Background
Caching using D2D communication in 5G is broadly categorized into three types. Synchronous caching, asynchronous caching and hybrid caching as depicted in Fig. 7. In synchronous caching, the BS maintains the information of all the mobile units connected. Moreover, it broadcasts messages in the network for each data entry to invalidate the caching. Also, no mobile unit is allowed to request the cache before the broadcast message. Asynchronous caching only broadcasts messages for the entry of
Review method
A systematic review for this survey paper was carried out by using the steps mentioned as follows.
- i.
Development of a review methodology
- ii.
Conducting the review, and
- iii.
Discussion of findings, analyzing and reporting the results.
Taxonomy of D2D content caching in 5G
The taxonomy of content caching techniques for D2D communication in 5G is as illustrated in Fig. 10. It includes the various parameters which categorize the available literature in cellular networks. The taxonomy contains four branches, each of which indicates different parameters, their sub-branches, and the corresponding literature. The parameters on which the classification is done are communication and connectivity of the mobile users, the type of storage used for caching, the quality of
Research challenges in D2D content caching techniques in 5G
In this section, we discuss various challenges while implementing wireless caching in D2D communication. Some of the issues identified in the survey include the speed of the mobile users, techniques to utilize the available resources, privacy and security, the need of big data technology for caching the content, and maintaining the quality of service to its end users. In Fig. 36, we represent the research challenges in D2D content caching techniques in 5G. Further, each issue has been discussed
Conclusion
The wireless data traffic is growing exponentially in recent years, with an increase in the number of smart connected devices. It results in the generation of a huge amount of data from these devices leading to issues such as storage, bandwidth, and higher data rates expectations during content sharing in 5G. The telecom operators are facing a challenge to fulfill these requirements of the users. In this paper, we reviewed content caching as an effective solution for D2D communication in 5G
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (166)
- et al.
A survey on 5G: The next generation of mobile communication
Phys. Commun.
(2016) - et al.
Lightweight and privacy-preserving RFID authentication scheme for distributed IoT infrastructure with secure localization services for smart city environment
Future Gener. Comput. Syst.
(2018) - et al.
Device discovery in LTE networks: A radio access perspective
Comput. Netw.
(2016) - et al.
Bayesian coalition game for contention-aware reliable data forwarding in vehicular mobile cloud
Future Gener. Comput. Syst.
(2015) - et al.
An intelligent approach for building a secure decentralized public key infrastructure in VANET
J. Comput. System Sci.
(2015) - et al.
Fog computing for Healthcare 4.0 environment: Opportunities and challenges
Comput. Electr. Eng.
(2018) - et al.
A survey on privacy protection in blockchain system
J. Netw. Comput. Appl.
(2019) - et al.
Cooperate Caching with multicast for mobile edge computing in 5G networks
- et al.
Joint computation offloading, resource allocation and content caching in cellular networks with mobile edge computing
- et al.
Cost-effective vehicular network planning with cache-enabled green roadside units