Let us take a look at the features of reactive systems:

• Responsiveness: Reactive systems respond in a timely manner. Such systems are meant to detect the problems quickly, and to deal with them efficiently. Reactive systems focus on rapid responses, and they usually set up an upper bound so as to provide the users with quick and consistent quality of service.
• Resilience: The systems are responsive in the event of a failure. The resiliency is achieved by replication, containment, isolation, and delegation. Failure of one system does not affect or impact the replicated system. The client either does not notice failure, or does not have to handle the failure as it is handled automatically.
• Elasticity: Primarily, the system stays responsive under different loads. A reactive system can scale up or scale down based on less load or more load. This also means that the system should have no bottlenecks. Scaling up or down can either be predictive or reactive based on the configuration with the combination of statistics by the reactive system.
• Message-driven: This property is perhaps the primary enabler to maintain the other properties of the reactive systems. A reactive system depends on asynchronous message communication. It ensures loose-coupling, location transparency, and clear isolation between message sender(s) and message receiver(s) / processor(s). Having explicit message-passing enables simpler load management, easier elasticity, and flow control based on message queuing mechanisms. Asynchronous non-blocking communication means that the recipient processors only consume resources while they are active, thus creating less system overhead.