GaN based electronic devices have progressed rapidly over the past decades and are nowadays starting to replace Si and classical III-V semiconductors in power electronics systems and high power RF amplifiers. AlGaN/GaN heterostructures have been, until recently, the materials system of choice for nitride based electronics. The limits of AlGaN/GaN technologies are now known and alternative routes able to overcome them are actively investigated. Nearly lattice-matched InAlN/GaN heterostructures have emerged as viable solutions for extending the frequency range achievable by GaN based devices. These heterostructures have also proven a very high thermal stability, becoming of high interest for extreme environment applications. However, despite those great potentials, InAlN/GaN based devices, in particular high electron mobility transistors, have suffered from high leakage currents, strong short channel effects and low breakdown voltage that made them not competitive with respect to AlGaN/GaN technologies. The goal of this thesis is to investigate, understand and control the mechanisms that limit the performance of InAlN/GaN based transistors. Particular attention in this thesis will be given to leakage currents, having gate or buffer origin. Concerning gate leakage currents, an accurate model for InAlN/GaN heterostructures will be established and the expected presence of deep levels will be confirmed by means of photocapacitance spectroscopy. Based on these results, it will be shown that two conduction mechanisms contribute to gate leakage currents. A first mechanism dominates in heterostructures with thin (<7 nm) InAlN barriers, while a second mechanism related to the appearance of degraded region becomes dominant in heterostructures with thicker barriers. Two approaches will be developed for the reduction of buffer leakage currents. A first approach consists in the growth of high quality heterostructures with extremely thin (50 nm) GaN buffer. It will be shown that thanks to this method high electron mobility transistors (HEMTs) with ON/OFF current ratios as high as 5x10^9 can be obtained. This approach will prove effective for the achievement of transistors able to display high performance at temperatures as high as 600 °C. A second approach, which will be developed for microwave transistors, makes use of heterostructures implementing an Al0.04Ga0.96N back-barrier . The reduction of ohmic contact resistivity will be also considered and the achievement of regrown ohmic contacts by NH3-MBE will be discussed. A last topic covered will be the thermal stability of InAlN based heterostructures at very high temperatures (850 °C). It will be shown that the stability can be greatly improved by appropriately capping InAlN with a thin GaN layer. This technique will be exploited for the achievement of state of the art InAlN/GaN heterostructures incorporating an in situ SiN passivation. To conclude, the performance of InAlN/GaN based transistors for both microwave and high temperature applications will be presented and discussed, demonstrating that these devices can now seriously compete with AlGaN/GaN based technologies.