LTE-Direct vs. WiFi-Direct as Enabling Technologies for Energy Efficient Machine-Type Communications over LTE-A Systems † , † , † , Jesus Alonso-Zarate * , † , and Antonio Iera † † University Mediterranea of Reggio Calabria, Italy, DIIES Department * Centre Technol` ogic de Telecommunications de Catalunya (CTTC), Barcelona, Spain Email: [antonio.iera|]@unirc.it; jesus.alonso@cttc.es Abstract—The enhancements introduced by Long Term Evolution-Advanced (LTE-A) cellular systems are mainly focused on boosting the data rates offered to terminals and to efficiently manage the transmission of large size data packets. However, focusing on the next-to-come scenarios of 5G wireless systems where machine-type communications (MTC) will compete in an equal footing with human-type communications (HTC), novel solutions are needed to efficiently exploit the radio resources for conveying the typical small MTC data. With this aim, this paper addresses the use of short-range device-to-device (D2D) communications as enabling technology to efficiently managing the radio spectrum and to reduce the energy consumption of MTC devices. We consider a scenario where MTC nodes are grouped in a cluster; among the cluster members, one terminal acts as aggregator in charge for: (i) receiving data from neighboring terminals via D2D links and (ii) relaying the aggregated data to the base station via macro-cellular link. The main contribution of this paper is to compare the performance of the most popular D2D technologies, i.e., WiFi-Direct and LTE-Direct, used to transmit data toward the aggregator. The performance evaluation in terms of latency, resource utilization and energy efficiency has been conducted in a wide set of scenarios by varying the number of clustered devices and their channel conditions. Index Terms—MTC; D2D; LTE-A; Energy Efficiency. I. I NTRODUCTION The support of machine-type communications (MTC) [?] [?] via cellular systems is mandatory to fulfill the requirements of future fifth generation (5G) wireless networks [?]. Indeed, MTC is expected to play a key role in the 5G scenario as testified by the exponential growth observed in the data traffic generated by heterogeneous devices (such as smart meters, signboards, cameras, remote sensors) which send their data to remote servers or to other machines (e.g., actuators) without (or with minimal) human intervention. This opens unprece- dented opportunities and business models [?] in different fields (e.g., transport and logistics, smart power grids, e-health, home and/or remote surveillance) belonging to the Internet of Things (IoT) ecosystem [?]. The effective provisioning of MTC over 3rd Generation Partnership Project (3GPP) Long Term Evolution-Advanced (LTE-A) system [?] represents one of the main challenges for cellular network providers due to the unique patterns of MTC traffic [?]. Indeed, MTC puts constraints in terms of: (i) energy efficiency of battery-powered machines, (ii) compu- tational efficiency of low-complexity embedded devices, (iii) low cost deployment to facilitate scaling, and (iv) low latency to support industry-compliant critical control applications [?]. Furthermore, MTC introduces a novel issue related to the fact that data transmitted by MTC devices are typically composed of few bytes. This is challenging because LTE-A systems are currently designed to efficiently guarantee high data rates to human-related traffic but suffer in terms of capacity when huge amount of MTC devices attempt to transmit few bytes in a very limited time interval [?][?]; this dictates for a paradigm shift on the packet scheduling as the minimum amount of radio resources that can be allocated to a single device in LTE-A could actually be too big for the actual needs of MTC traffic. According to above mentioned concerns, an energy-efficient transmission mode able to efficiently manage MTC traffic composed of few bytes simultaneously transmitted by huge number of devices still needs to be properly designed. With this aim, by following the trends presented in [?] which testified the meaningful benefits introduced by the exploitation of small-cells in MTC environments compared to traditional macro-cellular transmissions and in [?] which demonstrated the energy improvements achieved by sending small data with more robust modulation and coding scheme (MCS), in this paper we propose to handle MTC traffic through the use of network-assisted device-to-device (D2D) communications [?], whereby two or more devices in mutual proximity establish direct local links and bypass the base station (i.e., the eNodeB). The D2D technology is based on short range communications. This hypothesis is compliant with a large number of MTC ap- plications which foresee that the machines/devices are placed closely to each other. In our reference scenario, depicted in Fig. 1, MTC devices are grouped into different clusters and send their data to an aggregator via D2D links. Once collected the data by neighboring terminals, the aggregator relays the aggregated data (i.e., only one single data packet) to the eNodeB. The benefits offered by the presented scenario are in terms of: (i) better usage of the radio resources and (ii) low power trans-