QoS-Aware Stream Federation and Optimization Based on Service Composition

2. Motivation Scenario

Event services can be applied in different Smart City scenarios. In this paper, we study the scenario of smart urban travel planning, as regular citizens, developers and city stakeholders rank it highly¹. The city of Aarhus in Denmark has deployed a set of street-level traffic sensors. These sensors are paired as start nodes and end nodes. Each pair is capable of monitoring the average vehicle speed v and vehicle count n on a street segment (from the start node to the end node). Combined with the distance d between the two sensors, the estimated travel time t = d/v and congestion level c = n/d can be derived and published regularly as traffic report events.

Figure 1 shows some traffic sensor nodes (depicted as red dots) on the Aarhus city map. Suppose a user, Alice, has an important appointment in 15 minutes, and she has to travel from home (on segment A in Figure 1) to her work place (on segment F in Figure 1) within the time frame. Alice decides not to pick a route randomly since it is rush hour and there's a good chance that she may experience traffic congestion. Instead, Alice uses a travel planner application on her smartphone to select the fastest route. Alice would like to receive live traffic condition reports during her trip in case some traffic incidents happen on the selected route and a detour is necessary. To do that, she specifies the start and end location of the travel. She also wants to be sure that the time estimation is accurate, so she sets some non-functional constraints such as the accuracy of the estimated travel time above 90%.