• akka
• iot
• play
• reactive cloud
• scala

Management of network systems with hundreds to thousands of units may be a very challenging task. Network administrators and system integrators may need to know the health status of the network in real-time, to detect and prevent possible faults, to reduce labor costs and to simplify and speed up recurrent maintenance tasks and troubleshooting.
Current Network Management Systems (NMSs) only partially address all the issues described above. The majority of them are based on Simple Network Management Protocol (SNMP), which just offers basic operations with ap- plicability limited to a single Local Area Network (LAN).
In this thesis, we propose a cloud-based IoT approach to network and de- vice management, with the goal of overcoming the limitations of a standard SNMP-based NMS, providing a complete new set of features, such as geo- graphically distributed network remote control, bulk operations and predic- tive fault detection through Big Data analysis.
The proposed cloud service is designed and developed by using modern reac- tive Web technologies for Internet of Thinks (IoT) and Big Data, including Akka actor concurrency model, Play framework with Scala language, NoSQL MongoDB database and WebSocket protocol.
The result is a highly-scalable, non-blocking and responsive network man- agement system, relying on asynchronous and event-driven communication between components.
Tests for the cloud system are carried out automatically by using tools of- fered by Akka and Play frameworks. A simulated network unit is developed to simplify the testing phase and to evaluate the scalability of the cloud sys- tem by instantiating thousands of connected units.
Finally, the integration test with Fluidmesh Networks wireless mesh routers is performed and the correctness of the proposed communication protocol via WebSocket is evaluated.