Coexisting 802.11 and 802.15.4 single-cell wireless networks are experiencing significant performance degradation due to the aggressive nature of 802.11 (when compared to 802.15.4) and to their different traffic characteristics [1]. Providing tight delay guarantees to certain 802.15.4 applications (e.g., health monitoring with unsaturated, periodic traffic) is becoming increasingly infeasible, in the presence of coexisting WiFi with bursty and bandwidth-hungry traffic. Optimizing the performance of these coexisting networks (e.g., WiFi throughput maximization, while satisfying 802.15.4 deadlines) has been a challenging task, primarily due to the lack of: i) analytical models that take into consideration realistic network traffic conditions; and ii) accurate simulators for coexistence. In this paper, we address the aforementioned research challenges by modeling the transmission buffers of wireless devices as M/G/1 queues, and employ queuing theory and Markov Chain models to derive, for the first time, closed form solutions for throughput and delay in 802.11/802.15.4 coexisting networks. Using our proposed models, this paper presents a novel approach for joint MAC protocol tuning, that maximizes 802.11 throughput while satisfying delay constraints of 802.15.4. We validate our proposed solutions and models through new 802.11/802.15.4 coexistence capabilities in the ns-3 simulator (important for the research community).