Environmental energy is becoming a feasible alternative for many low-power systems, such as wireless sensor nodes. Designing an environmentally powered device faces several challenges: choosing the exact type of the energy harvester, the energy storage elements and determining the duty cycle of the application. With harvesting, the design process becomes even more difficult because it also has to take into account the unpredictability of the energy source. The contribution of this paper is a methodology that facilitates the analysis of energy harvesting nodes. The existing modeling strategies for battery powered systems are not suitable because they do not capture the uncertainty of the power source. Also, the metrics of interest for battery powered devices are different, as opposed to the harvesting powered ones: in the former case we search to maximize the system lifetime, while in the latter case a more expressive goal is to increase the system availability.