Budget-aware Video Crowdsourcing at the Cloud-enhanced Mobile Edge

Abstract : With convenient Internet access and ubiquitous high-quality sensors in end-user devices, a growing number of content consumers are engaging in the content creation process. Meanwhile, mobile edge computing (MEC) can provision distributed computing resources for local data processing. The MEC-enhanced video crowdsourcing application will gather user-generated video contents and collectively distribute them to the viewers of interest. To empower the crowdsourced video streaming at the edge, we investigate how to efficiently transmit data from content generators to the viewers. In particular, for a group of collaborative mobile users willing to share their data with the viewers, the content generation and delivery scheme is designed by considering the cost incurred by the crowdsourcing application. By leveraging the cloud resources available at the wireless base stations, the uploading or downloading server site is chosen for each user. To minimize the system makespan, i.e., the overall data transmission time among the generators and viewers, the user association scheme is also designed to efficiently utilize the diverse wireless radio resources. As compared with traditional centralized/distributed content delivery schemes, the proposed algorithm can improve the cost-effectiveness of distributed radio/cloud resources deployed at the mobile edge.

 EXISTING SYSTEM :

 ? We survey the existing research on caching for IoV with vehicles taking different roles, i.e., Vehicle as a Client and Vehicle as a Server. ? By analyzing existing traffic models, it was shown in that Vehicular Cloud computing is technologically feasible in dynamic scenarios, e.g., highways. ? Existing works rely on traffic surveillance cameras, which are not available on many roads, or GPS-based speed estimation, which only provides coarse estimates. ? Other existing studies on cloud-based SLAM for mobile robots, including DAvinCi and Rapyuta, also provide valuable lessons for developing edge-assisted SLAM for intelligent IoV. ? Existing studies on decentralized SLAM are mainly for mobile robotics.

 DISADVANTAGE :

 ? Edge-clouds can be helpful in two major problem classes: when there is no connection at all to a major infrastructure (usually the Internet) and when the current infrastructure is not enough to support all the traffic generated by the hosting of a large event. ? When we are talking about a network formed by mobile devices which have limited radio interfaces and power supply that communicate among each other via radio signals this problem becomes an important issue to tackle. ? Crowdsourcing is a distributed problem solving model in which a call to an undefined number of people is made in order to engage that same people in the process of solving a complex problem. ? Even though crowdsourcing is not fully adapted yet to the mobile phone workforce the widespread use of smartphones will reveal the full potential of this new problem solving approach.

 PROPOSED SYSTEM :

 • To extend the capabilities of VANET, the Internet of Vehicles (IoV) has been proposed to form a global network of vehicles, evoking collaborations between heterogeneous communication systems to provide reliable Internet services. • While Cloud-based SLAM algorithms have been proposed to alleviate the computation burden of vehicles, the propagation latency will not meet the real-time execution requirement. • Focusing on temporal information services in IoV, a distributed edge caching mechanism was proposed in based on the cooperation of RSUs and vehicles, in order to optimize both the temporal data and real-time requests. • In, edge caching according to the content size was proposed.

 ADVANTAGE :

 ? They compare the performance of FemtoCloud to an oracle which assumes accurate knowledge of all connectivity and execution time for every task on every device. ? The authors implemented three applications using the Honeybee framework to evaluate the performance and feasibility. ? Furthermore, smartphones’ multisensing capabilities as movement, geolocation, light, audio and visual sensors offer a variety of efficient new ways to collect data, enabling new crowdsourcing applications. ? They proved that the mechanism is computationally efficient, meaning that the winners and the payments can be computed in polynomial time (linear) which makes it scalable.