New generation mobile communication systems like Long Term Evolution (LTE) aim to deploy to customers a new mobile experience providing higher data rates and lower latencies that can make wireless devices a great platform to run a new whole set of services and applications, that was unimaginable just some years ago. However, the energy demands of battery powered devices, needed to support these new software applications substantially exceeds the capacity of the actual battery technology. The development of new architectures and procedures to build power-efficient and power-aware systems has become one of the main purposes in the design of new generation wireless networks. LTE exploits the idea of Discontinuous Reception (DRX) and Discontinuous Transmission (DTX), to provide a concrete solution to the power saving issue. The main point of this functionality makes the terminal to not continuously monitor control channels, allowing it to turn the radio frequency modem in sleep state for long periods, activating it only in well defined, suitable, instants. The objective of this thesis is to analyze the transmission of VoIP traffic over LTE networks trying to find the best solution to efficiently deliver data, fitting the DRX/DTX functionality to provide power saving and not perceptibly degrade the user experience. VoIP traffic has been studied with a bidirectional model, that emulates the interaction of speakers in a phone call. Two suitable scheduling policies for this kind of traffic, semi-persistent (SPS) and dynamic (DS), have been investigated underlining pros and cons of each solution. It has also introduced a simple control for SPS that can improve its performances. Other modeled entities include CQI reporting module, link adaptation algorithms, transmission and retransmission managers, a detailed implementation of a DRX/DTX framework and a simple solution to track power consumption statistics. The study involves the use of different configuration for channel quality reporting, link adaptation, DRX/DTX behavior with the extension of short DRX/DTX too. Every combination has been tested in single user, and multiple users scenario with 250 users all positioned at cell edge. The use of DRX/DTX functionality with SPS has been proved to provide good power saving effects due to the general synchronization between data source, SPS and DRX/DTX behaviors. This configuration generally induces acceptable delays in single user scenarios, but multiple users scenarios with too long DRX/DTX cycles can generate losses of about 4% of packets, and delays that propagate for the whole talk-spurt. Dynamic scheduling has shown to make good use of its channel quality tracking, performing an instantaneous correct estimation of needed resources. The cost of dynamic scheduling resides in its more frequent signaling activity, that impacts to a 33% higher power consumption compared to the SPS case. The common outcome of SPS and DS analysis is that the short DRX/DTX functionality does not bring further power savings to any of the studied scheduling solutions, due to VoIP traffic pattern.