As an effort toward the standardization, we further propose the potential universal architecture platform EWI (Embedded Wireless Interconnect), in wireless sensor networks, for replacing the existing OSI paradigm. EWI is built on two layers, which are Wireless Link layer and System layer, respectively. Studies from both experiential and theoretical perspectives are then considered for EWI.In wireless sensor networks, the optimization for lower energy consumption under application specific network QoS (Quality of Service) requirements, introduces a tremendous impetus to integrate multiple OSI (Open Systems Interconnect) layers. In the dissertation, we show that the optimization leads to cross layer design from the two ends, which are the Physical layer and the Application layer, respectively.Starting from the Application layer, we study LESOP (Low Energy Self Organizing Protocol) for target tracking in dense wireless sensor networks. The application QoS under consideration is the target tracking error, and a QoS knob is utilized to control the tradeoff between target tracking error and network energy consumption. Direct connections are found between the top Application layer and the bottom MAC (Medium Access Control)/Physical layers. Moreover, unlike the classical OSI paradigm of communication networks, Transport and Network layers are excluded in LESOP to simplify the protocol stack.Starting from the Physical layer, we first concentrate on the probability of successful radio packet delivery. By exploiting the tradeoff between this probability and network energy consumption, CTP-SN (Cooperative Transmission Protocol for Sensor Networks) shows that the sensor nodes cooperative radio transmission reduces the outage probability exponentially when the nodes density increases. In MSSN (Sensor Networks with Mobile Sinks), on the other hand, the probability of successful information retrieval on the mobile sink is under consideration. Optimal and suboptimal transmission scheduling algorithms are then studied for MSSN, by exploiting the particular tradeoff. In both studies, optimizations lead to compound Link layer and Physical layer design.