An improved virtualization layer to support distribution of multimedia contents in pervasive social applications

Abstract

Pervasive social computing is a new paradigm of computer science that aims to facilitate the realization of activities in whichever context, with the aid of information devices and considering social relations between users. This vision requires means to support the shared experiences by harnessing the communication and computing capabilities of the connected devices, relying on direct or hop-by-hop communications among people who happen to be close to each other. In this paper, we present an approach to turn mobile ad-hoc networks (MANETs) into stable communication environments for pervasive social applications. The proposal is based on an evolution of the VNLayer, a virtualization layer that defined procedures for mobile devices to collaboratively emulate an infrastructure of stationary virtual nodes. We refine the VNLayer procedures and introduce new ones to increase the reliability and the responsiveness of the virtual nodes, which serves to boost the performance of routing with a virtualized version of the well-known AODV algorithm. We prove the advantages of the resulting routing scheme by means of simulation experiments and measurements on a real deployment of an application for immersive and collective learning about History in museums and their surroundings.

Introduction

The steady advances in the realm of pervasive computing favored by the ever-growing popularity of smartphones, together with the new interaction patterns propelled by the social web, pave the way to a new era of information services tailored to the people׳s physical and social context (Drego et al., 2007, Zhou et al., 2012). Researchers in the area of pervasive social computing envisage many opportunities to enable meaningful interactions among groups of nearby people —either acquaintances or strangers— to help them make the most of their environment, ranging from the orchestration of activities in venues or events that attract people with potentially-related interests (e.g. museums, concert halls or campsites) to advances in the vision of smart cities, related to the planning of personal mobility or the celebration of location-based urban games (Schuster et al., 2013). As explained in Ben Mokhtar and Capra (2009), while pervasive computing aimed to facilitate the realization of users׳ daily tasks by changing the way they individually interact with the physical environment, pervasive social computing aims to facilitate the realization of those tasks that need to consider social relations between users. While the former placed a strong emphasis on the self component, the latter emphasises the social component. While web 2.0 websites enabled online interactions, pervasive social computing aims to change the way individuals interact with co-located people.

From the technological point of view, realizing the concept of pervasive social computing requires means to support the shared experiences by harnessing the communication and computing capabilities of the connected devices as an integrated whole. At the lowest level, due to physical proximity, the devices should be able to exchange data packets directly or within few hops in an ad-hoc network—rather than sending them out to remote Internet servers that would merely echo the same packets downlink—and harness the strengths of peer-to-peer (P2P), opportunistic networking (Sun, 2001, Conti and Giordano, 2013). The ad-hoc networks should serve as a channel to distribute a wealth of multmedia contents generated locally or downloaded from the Internet through any WiFi, DSRC, WiMAX, 3G or LTE connections available to the mobile devices. In doing so, it is necessary to deal with the challenges raised by the properties of the wireless medium, the varying topologies that emerge from the users׳ movements, and the fact that the devices may have limited memory, computational power and battery.

In this paper, we introduce an approach to support pervasive social applications by turning the mobile ad-hoc networks (MANETs) into more stable communication environments. Our proposal is based on an evolution of the virtualization layer presented in Dolev et al. (2004) and Brown et al. (2007) (called the Virtual Node Layer, VNLayer), which put forward procedures to create an infrastructure of stationary virtual nodes to ease the routing problem and the maintenance of persistent state information in the area covered by an ad-hoc network, notwithstanding the movements of the real nodes. Brown et al. (2007) discussed the advantages of the programming abstraction enabled by the VNLayer, whereas Wu et al. (2011) and Patil and Shah (2012) separately proved that a virtualized version of the AODV routing algorithm (Perkins et al., 2003) (called VNAODV) can outperform AODV itself in terms of route stability and packet delivery ratios. Later on, Wu proved the advantages of virtualization for another routing algorithm (RIP, Malkin, 2000) and ancillary protocols like DHCP (Wu, 2011). Here, we present a number of refinements to both the VNLayer and VNAODV, leading to new versions we have called VNLayer+ and VNAODV+. By means of simulation experiments and measurements on a real deployment, we prove that this solution achieves better performance than others in supporting a pervasive social application for History-related museums called REENACT (previously described in López-Nores et al., 2013a, Blanco-Fernández et al., 2014). Specifically, we have compared the performance achieved by the REENACT app in the distribution of multimedia contents with the five routing solutions shown in Fig. 1: plain AODV, VNAODV on top of the VNLayer, VNAODV+ on top of the VNLayer+, OLSR (Clausen et al., 2003) and ARA (Guenes et al., 2002). AODV, OLSR and ARA are the most representative examples of routing algorithms used for P2P distribution of contents in MANETs, along with DSDV (Perkins and Bhagwat, 1994) (an antecedent of OLSR) and Bee (Wedde and Farooq, 2005) (conceptually similar to ARA) (Gurumurthy, 2009, Tang et al., 2005, Sbai et al., 2010, Hwang and Hoh, 2009, Dhurandher et al., 2009, Castro et al., 2010, Barbeau, 2012, Wang et al., 2014).

The paper is organized as follows. First of all, Section 2 includes an overview of routing protocols for MANETs. Then, we present the main procedures of the VNLayer (Section 3) and details of how AODV was adapted to work on top of it (Section 4). The refinements we have included in the VNLayer+ and in VNAODV+ are presented in Sections 5 and 6, respectively. Section 7 presents the REENACT application, the scenarios of our measurements and simulations and the comparison results. Finally, conclusions are given in Section 8 along with the motivation of our ongoing work.