A publish-subscribe based framework for event notification in vehicular environments

Abstract

In this work, we consider a vehicular environment in which vehicles moving along roads in a city are interested in different types of information or events generated at different parts of the city. We investigate the use of the publish-subscribe framework for such environments in which vehicles subscribe to a service provider for specific types of events through roadside units; the events are also reported to the service provider. The service provider delivers the events to the subscribed vehicles within the validity periods of both the subscriptions and the events through roadside units placed along the trajectory of a vehicle. We propose a problem called the Minimum Cost Event Placement Problem which addresses the issue of delivering all the events to all the subscribed vehicles within the validity periods while incurring minimum cost of utilization of the roadside units. We first prove that the problem is NP-complete. We then propose three heuristic algorithms to solve the problem. A lower bound on the cost of utilization of the roadside units for a given set of subscriptions and events is presented next. Detailed simulation results in different city traffic scenarios are presented to show that the costs achieved by two of the heuristics proposed are close to the lower bound proved.

Mobility Aware Event Dissemination in VANET

“…We refer to this problem as the Restricted-MinCost-EPP problem, and can then be formally defined as follows [10]: The problem has been proved to be NP-complete [10]. In the next section, we will describe an approximation algorithm for the offline version of the problem.…”

“…Mukherjee et al [10] define the Minimum Cost Event Placement Problem, and prove that the problem is NP-complete. They propose three greedy heuristics for the offline version of the problem.…”

“…The function θue gives the total time for which the event e needs to be broadcast at u, and captures the overlap between time intervals of vehicles passing u whose subscriptions can all be satisfied with e. For example, if Fue = {f1, f2, f3}, with the time intervals of [5,10], [12,18], and [14,20] respectively, then the three subscriptions can be satisfied by broadcasting e from u for the duration [5,10] and [12,20]. The function θue will return (10 − 5) + (20 − 12) = 13 in this case.…”

“…In this section, we state the MinCost-EPP problem formally as defined in [10]. Let E denote the set of events reported to the SP, and F denote the set of subscriptions issued from vehicles in the city.…”

“…In such a scenario, it is important for the SP to decide when, where, and for how long to place the events in different RSUs such that the vehicles can receive notifications for the events they are subscribed for within the validity periods of both the subscriptions and the events while incurring minimum cost to the SP. The problem is referred as the Minimum Cost Event Placement Problem (MinCost-EPP) [10].…”