This application should be a working prototype of a system that aims to manage a smart parking area. The system is virtual in the sense that the components (sensors, cars, etc) are not real but are software components that simulate a plausible behavior.
The main elements that compose the system are:
-
an environment: it is where the park area is placed and everything takes place. It is, for simplicity, a grid-based world, that is an (n x m) matrix where each cell could be a street block or a park place.
-
a sensors layer: each car parking place is equipped with a parking sensor able to determine if the place is free or busy. Each sensor send this information to a central server. It also provide a (simulated) local area network that could be used by a client application to retrieve information in case the server is not available.
-
a server application: it knows everything about the parking area but it may suddenly break down. The server receives information from sensors and uses them to produce a dynamic park map (for eg. a graph). The server interacts also with drivers (client applications) sending the current map data that will be used to locate the nearest parking place.
-
some client applications: that is drivers that would like to park their car or to locate a parked one. The client application interacts with the server, if available, to retrieve an updated park map and find the nearest parking place available. If the server is offline, it tries to interact with the nearest sensors to retrieve some useful information.
The final system should provide the following functionalities:
-
find the nearest free park place
-
locate the parked car
-
...
The above functionalities should be supported in a client-server interaction schema if the server is available or using local interactions with near sensors otherwise.
The system should face the following aspects:
-
it is composed by a dynamic and possibly large set of components. At a given moment the number of clients is unpredictable
-
it is open, that is components could be added or removed while the system is working
-
it is heterogeneous, clients, server and sensors application may be written in different programming language
-
it should work in a centralized and decentralized/distributed way
-
the interactions are asynchronous and loose coupled
-
the components are situated in a (virtual) world
How to deal with the previous aspects?
-
strong modularity, rigid boundaries definition
-
loose coupled interactions, eg. MOM adoption to enable flexible communication among different aplicatons and standard data formats (JSON)
-
explicit messages and protocols definition
-
few assumptions between components, few dependencies
-
Install and start RabbitMQ services
-
Download the source
-
Import the projects in Eclipse
-
Run gradle eclipse in isac.client, isac.server, isac.environment to set up dependencies
-
Run gradle build in isac.client, isac.server, isac.environment if you want to create a runnable jar
At the moment components interactions are implemented using RabbitMQ.