Ensuring the sustainability of real-time embedded system under both QoS and Energy Constraints

Nowadays, wireless sensor networks (WSNs) are more and more used in applications such as environment monitoring, healthcare monitoring, etc...The challenge in sensor networks is to ensure the sustainability of the system by guaranteeing the required performance level. However, with the limited capacity of finite power sources and the need of guaranteeing a long lifetime of those systems, it is suitable to use energy harvesting which allows to supply low-power electronic systems by converting ambient energy into electric power. Hence, our study is concerned with the problem of soft periodic and aperiodic tasks scheduling in sensor nodes powered by energy harvesters. In this paper, we address this issue by proposing three energy-aware schedulers, namely BG-Green-RTO, BG-Green-BWP and Green-AWP which aim to improve the responsiveness of aperiodic tasks while still guaranteeing the execution of periodic tasks considering their timing and energy constraints. Such algorithms allow to gracefully cope with processing overload and energy starvation. Moreover, a simulation study permits to show their performance. 1Introduction Wireless sensor networks have taken more and more place in diverse application domains. For example, in agriculture, it could be possible to deploy a wireless embedded network to detect various measures across a field such as temperature, light levels, soil moisture, ... This example implies that measures should be collected at a specific time and should be treated on time to react correctly to a given situation. Majority of wireless sensor networks represent real-time embedded systems. A real-time system is a system which must guarantee a response within a specified time (often referred to as "deadline"). In other sense, it is a system in which the correctness depends not only on the logical result but also on the time in which it was delivered. So that, if the time constraints are not respected it may lead to a system failure. According to their time constraints, real-time systems are classified into three categories: hard, soft and firm [6]. In a hard real-time system, all deadlines must be met imperatively. In a soft real-time system, missing deadlines will only cause a performance degradation without leading to a system failure. A firm real-time system is a specific case of a soft real-time system so that it must meet its deadlines but with a degree of flexibility. Batteries are the source of power for most embedded system applications. However a battery has a finite lifespan, limited energy density and capacity so when all its energy is consumed, the sensor must be retrieved to replace the battery. Nevertheless, It is not as simple to replace it, especially when the sensor is deployed in an inaccessible place where any human intervention may be either costly or impractical. Thus, such systems should be designed so as to ensure a continuous functioning without having a periodical maintenance due to replacing or recharging batteries. Renewable energy sources are available in an unlimited quantity and there is a variety of available techniques for energy harvesting (i.e. the process of extracting energy from the surrounding environment) such as solar, piezoelectricity, o