Hybrid multicast-unicast streaming over mobile networks

Abstract

Modern cellular networks support unicast and multicast video streaming. Streaming over unicast results in high energy saving for mobile devices, but it can only serve a limited number of users.Whereas multicast streaming can serve large number of mobile users, but it increases the energy consumption for mobile devices. We formulate a resource allocation problem that considers both unicast and multicast simultaneously to maximize the average energy saving across all mobile devices and to serve large number of mobile users. We propose efficient algorithms to solve this problem. Our simulation results show that the proposed algorithms: (i) result in high energy saving close to optimal solutions, (ii) serve large number of mobile users compared to that of unicast-only approach, (iii) are more elastic than unicast-and multicast-only approaches, and (iv) lead to higher energy saving as more network resources are allocated to streaming services.

Dynamic configuration of single frequency networks in mobile streaming systems

“…Our proposed resource allocation algorithm in this paper aims at increasing the bandwidth utilization of a cellular network, but it is different from the works in [15][16][17] in two main aspects: a) it utilizes an adaptive and flexible scheduling process, and b) it takes an advantage of three transmission modes. The fraction of radio resources reserved for multicast services is assumed to be constant in most existing scheduling algorithms, including those schemes in [11][12][13][14][15][16][17]. In our case, the resource distribution between unicast and multicast connections is done dynamically based on which served request leads to better spectral efficiency.…”

“…Although a number of algorithms have addressed the transmission scheduling problem for multicast services [11][12][13][14], a few research efforts have considered the hybrid unicast-multicast approach in the allocation problem. For example, Monserrat et al [15] and Lee et al [16] present two schemes in which both unicast and multicast connections are served with an objective of reducing the service blocking probability, whereas Rahman et al [17] utilize the hybrid approach to minimize the average power consumption of mobile terminals. Our proposed resource allocation algorithm in this paper aims at increasing the bandwidth utilization of a cellular network, but it is different from the works in [15][16][17] in two main aspects: a) it utilizes an adaptive and flexible scheduling process, and b) it takes an advantage of three transmission modes.…”

“…For example, Monserrat et al [15] and Lee et al [16] present two schemes in which both unicast and multicast connections are served with an objective of reducing the service blocking probability, whereas Rahman et al [17] utilize the hybrid approach to minimize the average power consumption of mobile terminals. Our proposed resource allocation algorithm in this paper aims at increasing the bandwidth utilization of a cellular network, but it is different from the works in [15][16][17] in two main aspects: a) it utilizes an adaptive and flexible scheduling process, and b) it takes an advantage of three transmission modes. The fraction of radio resources reserved for multicast services is assumed to be constant in most existing scheduling algorithms, including those schemes in [11][12][13][14][15][16][17].…”

A joint multicast/D2D learning-based approach to LTE traffic offloading

“…UEs have no assurance of reception because the radio channel could suddenly degrade during data reception (e.g., due to fast fading or mobility). For this reason, mechanisms similar to panic retransmission are considered in several works in the literature [27,21]. presence of UEs experiencing severe channel conditions, the eNB emits at a modulation that leaves them in outage.…”