Leveraging on Electric Vehicles for Big Data Transfers

Both governments and car manufacturers have recently made huge investments in electric vehicles (EVs) to turn them into a viable solution to cope with problems such as pollution due to internal combustion engines and increasing gas prices. For instance, the Renault-Nissan Alliance has invested 4 billion in EVs since 2010. In particular, they signed a partnership with Better Place, a company that operates battery swapping stations in both Israel and Denmark. At the same time, 2.8 Zettabytes of content were created or replicated in 2012, and is forecasted to explode to 40 ZB by 2020, according to IDC [1]. Part of these numbers includes data generated by applications that can tolerate delivery delay in the order of hours or days (e.g., virtual machine migration in cloud architectures or content replication among a content provider network). In this paper, we advocate the opportunistic use of electric vehicles as carriers for delay-tolerant big data transfers. In this scheme, EVs are equipped with one or more removable memory storage devices attached to the battery module. Vehicles use battery swapping stations as offloading points (OPs), where data are either loaded onto the vehicle, switched from one vehicle to another, or unloaded from the vehicle into the Internet. The decision on whether to load or not load data on the vehicles is made at the offloading points, depending on the destination of the data combined with the traffic expected on the roads adjacent to the OP. In a companion work, we focused on showing the feasibility of this scheme by considering a single road segment using the average daily traffic [2]. Here, we introduce a new formulation of the problem. We raise new challenges by taking into account new elements such as routing and data losses occurring when EVs show unexpected behaviors between two OPs (data leakage). As shown in Figure 1, we consider an overlay to formally define our framework, which is composed of the underlaying road infrastructure (layer 1) and the network of offloading points (layer 2). We denote the underlaying road infrastructure by a directed graph GR = (NR, LR) where NR refers to road junctions and swapping stations (nodes in the graph) and LR refers to roads (links in the graph). We denote Haifa Source Destination