1 Social-aware Content Delivery with D2D Communications Support for Emergency Scenarios in 5G Systems Antonino Orsino, Leonardo Militano, Giuseppe Araniti, Antonio Iera ARTS Lab., DIIES Dept., University Mediterranea of Reggio Calabria, Italy e-mail:{antonino.orsino, leonardo.militano, araniti, antonio.iera}@unirc.it Abstract—In future Fifth Generation (5G) cellular systems, mobile social networks are deemed to play an important role in the information and multimedia content dissemination. From the technological perspective, Device-to-Device (D2D) communi- cations and multicast services have been identified as potential enabling technologies for a fast and efficient content transmis- sions. In this paper, we propose a viral information diffusion approach that brings together these different ingredients with the objective of minimizing the content delivery time experienced by the users over the cellular infrastructure. To this aim, we model an expected diffusion time where network metrics and social relationships are combined and used to drive the choices in the dissemination process. In particular, a new metric called social network contact time is introduced, able to efficiently characterize the user behavior in checking a notification received on a smart device. A novel solution effectively combining multicast, D2D communications and social network concepts is proposed and evaluated through simulations in a wide range of scenarios. The comparison with the legacy LTE multicast technology shows how the proposed solution is able to introduce considerable gains in terms of (i) average diffusion time per user, (ii) average throughput, and (iii) Jain’s fairness index. Index Terms—Content Delivery, LTE-A, 5G, D2D, Multicast, Mobile Social Networks. I. I NTRODUCTION T HE fast deployment of new communication technologies, enhanced devices, and multimedia services will lead to a drastic change in the way wireless systems are used. In the future 5G systems vision, it is expected that current human-centric communications will be flanked by a tremen- dous increase in the number of small devices represented by machines, or, more generally, “small” objects. Each person will be constantly connected not only with other “humans” but also with the objects of interest in order to make everyday life more efficient, comfortable, and safe [1]. In this vision, the adoption of Long Term Evolution (LTE) radio technology will act as a lifeline, in all situations including times of a more general emergency. Therefore, in addition to a more robust communication, future 5G system should provide low latencies and timely connectivity in case of natural disasters such as earthquakes, tsunamis, floods, hurricanes, etc. On the other hand, it has been shown in recent tragic events that social networks and social media may be a fundamen- tal support for disseminating information to the authorities control center by also alerting messages to the end-users involved in the disastrous event. This means that the devices disseminated in a given environment jointly with the users may become source of crucial information that may support the definition of an overview of the current situation in the incident/emergency area. Context-awareness can further benefit from objects building the Internet of Things (IoT) [3], [2] and not necessarily belonging to a specific end-user [4]. In particular, the avalanche of data that can be collected by devices and humans should be efficiently elaborated to help, with crucial and timely information, rescue teams, police, or single end-users in life-threatening situations. Concerning the information diffusion, and, as a conse- quence, the efficient collection of data from the environment, users’ social ties [5] and Device-to-Device (D2D) commu- nications are proposed as effective paradigms to reduce the diffusion time [12]. Direct communication between two (or more) nearby User Equipments (UEs) in mutual proximity is considered as a promising technology to be used in future 5G system. The benefits D2D introduces are in terms of, among others, coverage extension, improved spectrum utilization, higher throughput, and lower energy consumption [6], [7]. Due to these enhancements, it is expected that proximity- based communications may be useful for context-aware data collection and information diffusion in emergency situations when data has to be sent from the area of interest to the central base station in both uplink and downlink directions through reliable and low latency links. In this paper, we propose a framework where context- aware information is collected from “smart” objects and users randomly scattered in a certain area of interest. In particular, once all the relevant data are collected, the base station elaborates the information by distributing them to the end- users to minimize the information diffusion time. To this end, proximity-based communications are exploited both in uplink and downlink to investigate the potential benefits and enhancement introduced by this paradigm. A further main contribution of our work is the definition of an information diffusion approach based on the intertwined contribution of typical LTE-A transmission modes and social interactions among users and their respective social network platforms. To this aim, the proposed solution exploits a new social metric named social inter-contact time (SICT), which represents the frequency of interaction of an end-user with a social network platform it is registered to. A simulation campaign under a wide set of parameters settings will demonstrate the