With the successful deployment of IEEE 802.11a/b/g wireless local area networks (WLANs) and the pent-up market demand for high speed wireless access, the IEEE 802.11n protocol has emerged as a promising standard with major enhancements. In order to scale to the high physical layer transmission rate of up to 600 Mbps, the current medium access control (MAC) protocol requires significant improvements. Towards this end, several packet aggregation schemes have been proposed to reduce the inherent protocol overhead. However, choosing large frame sizes would result in increased channel access delay and jitter especially in delay-critical applications. In this paper, we propose an analytical model to obtain the optimal frame size with regard to the delay constraints of nodes. In particular, we model IEEE 802.11n aggregation scheme as a constrained convex optimization problem so as to maximize network throughput performance while considering delay constraints. Simulation results justify that our proposed model exactly satisfies average delay requirements of network nodes.