Green content sharing mode: D2D Coordination Multiple Points Transmission

Highlights

• Propose D2D Coordination Multiple Points Transmission.

• A distributed algorithm for matching providers to demanders.

• A packet split algorithm to achieve load balance among providers.

• Model the file reconstruction problem of collaboration demanders as the shortest Hamiltonian path problem.

• A best-effort distributed greedy algorithm framework to solve the shortest Hamiltonian path problem.

Abstract

In next-generation 5G cellular networks, device-to-device (D2D) communication has emerged as an effective solution to support the growing popularity of multimedia contents for local service. Conventionally, the D2D content sharing mode is “one-to-one” matching, i.e, one demander will select one provider to request files from it. Under this mode, it is hard to cope with the growing demand for multimedia services for mobile users due to limited battery capacity for mobile devices. In this work, we propose an energy-efficient content sharing system via a novel D2D Coordination Multiple Points Transmission (D2D-CoMP), which shares content among multiple users to reduce the power consumption per user device. The highlights of this work lie in three parts. Firstly, the strategy for matching providers to demanders subject to self-interference constraints is formulated as a classical maximum weighted matching problem, in which the optimal solution can be derived when network-wide information is known, and also an effective distributed algorithm. Secondly, we design an optimal packet split algorithm for D2D-CoMP under comprehensive consideration of two aspects of communication efficiency and energy consumption to solve the problem how many data packets each provider transfers. Thirdly, we model the file reconstruction problem of collaboration demanders as the shortest Hamiltonian path problem and illustrate the file reconstruction process. Moreover, we develop a best-effort distributed greedy algorithm framework to find the shortest file reconstruction path. Importantly, numerical results demonstrate that the proposed mechanism can greatly reduce the energy consumption of each device with little or no increase in transmission delay.

Introduction

The recent widespread use of mobile Internet complemented by the advent of many smart applications has led to an explosive growth in mobile data traffic over the last few years. This results in that the demand for bandwidth in cellular networks is expected to surge exponentially over the next several years [1], [2]. However, traditional ways of improving throughput within the context of cellular networks have suffered from two major challenges. Firstly, increasing the physical-layer capacity of wireless links becomes difficult, since the physical-layer technologies used in 4G/LTE (Long Term Evolution) networks, such as MIMO-OFDM (Multiple Input Multiple Output-Orthogonal Frequency Division Multiplexing) with capacity-achieving codes and interference coordination, have approached their theoretical limits. Secondly, decreasing the cell size by using femto-cell networks is a viable, yet costly, approach for improving throughput, as the cost of providing backhaul connectivity of these small base stations needs to be taken into account [3], [4].

Device-to-device (D2D) communication, as a communication method in 5G has been widely treated as a promising solution, mainly due to its unique advantages of offloading cellular traffic, better system throughput, higher energy efficiency and robustness to infrastructure failures. Essentially, D2D communications refer to the devices communicate with each other directly without an infrastructure of access point (e.g., base station) [5], [6], [7], [8]. Conventionally, a device exclusively relies on the cellular communication to retrieve the content it desires. With D2D communication, however, if the same content is available in the vicinity of the device, the content can be directly retrieved from the neighbouring devices [9], [10].

In such a D2D content sharing scenario, a user who wants a certain content is referred to the demander, and a user who has the desired content and does temporally not retrieve contents for himself is called the potential content provider. The existing academic literatures on such D2D content sharing scenario (e.g., [6], [9], [10]) have largely focused on how D2D communication can be used effectively to offload cellular traffic from the Internet, as devices request different pieces of content over cellular networks. What they have in common is to obtain performance gain and offload gain depending on selectively caching popular content locally and optimal matching for demander–provider pairs. However, there exists an issue that has not been paid attention to in these existing literature. Actually, the potential content providers with good link quality and many social ties will be likely selected to provide content to demanders frequently. This will put a great burden on such providers, especially the energy consumption. Moreover, due to the social similarity of users [7], [9], [11], [12], [13], if a device requires a file at a certain time, there will be a great possibility that the other devices in its vicinity also require the same file. Consequently, in this paper, we wish to study how D2D communication can be used to offload cellular traffic from Internet by the collaborative multiple devices transmission in a green way, as devices request files over cellular networks. We call this transmission as D2D Coordination Multiple Points Transmission (D2D-CoMP). This D2D content sharing scenario in a single-cell cellular network is illustrated in Fig. 1. In this figure, if user 1 wants to request a file, she can cooperate with her neighbour user 2 to request the file. In this case, she can get half of the file from user 3 while user 2 acquires the other half of the file from user 4. Then they share content with each other to get the complete file. In this content sharing process, user 3 and user 4 only transmit a half of the file, so they consume less energy than that for transmitting a complete file. Unlike the existing works about D2D cooperative content sharing, our D2D-CoMP focuses more on energy conservation. Before the link is established, we try our best to include demanders with the same content requirements into the communication, and obtain the content from as many providers as possible, thus reducing the burden of a single provider. Therefore, there are many differences between our D2D-CoMP and other D2D content sharing schemes. In Table 1, we made a summary. As such, the following three fundamental questions should be systematically answered:

Question 1: How to achieve the optimal matching among demanders and providers of contents?

It is general that only part of providers can come into the sight of the demand, while several demanders may simultaneously look forward to the help from the same provider. Whether by the limited storage and transmission capability of devices or by their various requests, it seems impossible for any potential provider to help all demanders freely. Hence, the core of matching demanders to potential providers is to choose and coordinate the providers from candidates for each demander. However, it is difficult to formulate and address this issue. In [17], the authors formulated this problems as a maximum weighted matching problem with D2D link rates as the weights assigned to each pair. In [18], a hypergraph based three-dimensional matching problem can be formulated by considering the content distribution, the pairing between demanders and potential providers, and the reuse of cellular resources for content sharing links. But, in these works, the matching solution can only be found in a centralized fashion because the global information is needed. It will generate extra traffic between the devices and the base station, adding additional burden to the cellular network. In [10], a distributed algorithm is desirable to solve the problem locally, without involving the base station. In order to achieve the optimal matching without increasing the burden of base station, we try to design a distributed optimal matching algorithm with D2D link rates as the weights assigned to each pair.

Question 2: How to achieve the equitable and green D2D transmission among demanders and providers?

In our proposed D2D-CoMP, multiple providers collaboratively transfer data packets to demanders. But, how many packets each provider transfers is a hard issue. As it is known to all, the condition of each provider is different at any time, some has high residual electricity, some has high link rate, and so on. Moreover, how to guarantee the load balance [19], [20] among devices is also a tough issue. Some nodes will take on high load transmission tasks for a long time and consume electricity too much and turn off to become dead nodes. In [21], [22], the authors study the task offloading in Ultra-dense Network and this enlighten us. Hence, we may consider the condition of each provider to determine how many packets it can transmit.

Question 3: How to achieve the efficient and reliable file reception?

Because many users request a file together and each receiving user receives only a small part of the file, so the users are required to share data with each other, so that each requester can get a complete file. How to share files between users, what kind of communication links need to be established can efficiently ensure that each user receives a complete file? This is a difficult issue to address. Aim at this problem, we try to find the shortest file reconstruction path by means of the related theory of Hamiltonian path [23], [24] in graph theory.

Motivated by the above observations, we introduce such integration into D2D-CoMP scenarios. The main contributions are as follows:

(1) With respect to matching providers to demanders, we formulate it as a maximum weighted matching problem, and D2D link rates are weighted to each sender–receiver pair. Though the Hungarian algorithm [18] can be used to solve the matching problem, it is centralized in nature and requires the global information. In order to solve this intractable problem with low complexity and high accuracy, we propose an asynchronous and distributed algorithm based on [25] to solve this problem.

(2) We investigate the optimal packet split algorithm for D2D-CoMP under the comprehensive consideration of two aspects: communication efficiency and energy consumption. Its advantage is to achieve load balance between multiple devices.

(3) We model the file reconstruction problem of collaboration demanders as the shortest Hamiltonian path problem, with the weights based on the link rate of each link between demanders. In order to solve this NP-hard problem with low complexity and high accuracy, we develop a best-effort distributed algorithm framework to satisfy needs.

The rest of this work is organized as follows. Section 2 presents the system model and problem overview. In Section 3, we show our decentralized solutions to the matching problems between providers and demanders. We propose the Optimal Packet Split Algorithm in Section 4. In Section 5, the file reconstruction problem of collaboration demanders is modelled as the shortest Hamiltonian path problem and a best-effort distributed algorithm is designed. We demonstrate the effectiveness of our algorithms through simulations in Section 6. Finally, the extensions and implementation details of our algorithms are discussed in Section 7, and the summary and future work are concluded in Section 8.