Data partitioning and scheduling schemes for federated platoon-based vehicular cloud

The organization of vehicles into platoons is very promising due to its contributions to intelligent transportation systems, and the reduction in traffic congestion and fuel consumption. As the powerful processing units and other resources embedded in vehicles may not be fully utilized during the entire travel period of vehicular platoons, we claim that the vehicles can collaborate as a single unit to form a federated platoon-based vehicular cloud to meet the high demand for computing resources and services in vehicular environments. In order to make advancements in the deployment of federated platoon-based vehicular cloud, data partitioning and scheduling schemes that distribute data chunks of large and divisible application data among platoon vehicles are proposed considering the characteristics of vehicular resources, network parameters, and the position of vehicles in the platoon. The data partitioning and scheduling schemes, which are modeled based on the different information flow topologies of vehicular platoon, consist of the Bi-Directional Recursive (BD-R), Bi-Directional Interlaced (BD-I), Bi-Directional Lead Recursive (BDL-R), Bi-Directional Lead Interlaced (BDL-I), Bi-Directional Lead Aggregate Recursive (BDLA-R) and the Bi-Directional Lead Aggregate Interlaced (BDLA-I). Performance analysis carried out through realistic simulations showed that while the BDL-R and BDL-I schemes have the best performance in terms of task execution time, the other schemes have the advantage of enforcing priority in unprocessed data transmission and dependency in the aggregation of processed data chunks by each platoon member. Analysis of the impact of the proposed data partitioning and scheduling schemes on platoon string stability will be examined in future studies.