Fragmentation and AES Encryption Overhead in Very High-Speed Wireless LANs

Abstract

In this paper, we study the overhead introduced by the advanced encryption standard cipher in the context of wireless LANs, specifically at the medium access control layer, as described in the 802.11 standard developed by the 802.11n task group. We compute the maximum throughput, optimal frame, and fragment sizes which can be achieved in this context and compare them to the optimal values when encryption is not used. I. INTRODUCTIONNE of the main challenges in wireless LANs (WLANs) nowadays is to develop a medium access control (MAC) layer that will not decrease the efficiency of the MAC layer when PHY rates are increased since as studied by Xiao et al. in [17,18,14], a theoretical throughput upper limit exists, indicating that by simply increasing the data rate without reducing overhead, the enhanced performance, in terms of throughput and delay, is bounded even when the data rate goes into infinitely high. Of the existing models, we are particularly interested in Aggregation with Fragment Retransmission (AFR) scheme, which was initially proposed in the IEEE 802.11n task group [9], and then developed more comprehensively in [6]. In this work, multiple frames are aggregated into a larger frame before being transmitted to the physical layer (PHY). If the size of a frame is larger than a pre-established threshold, the frame is divided into fragments before being aggregated. Transmission errors are handled by retransmitting only the fragments of the frame that had been corrupted.However, the work in [6] does not consider security, i.e., encryption algorithm AES, which is used in IEEE 802.11i. In other words, when IEEE 802.11n and IEEE 802.11i are both adopted, AES over the high speed wireless LANs (WLANs) must be considered. With this motivation, in this paper, we analyze the overhead introduced by AES, when added to the aggregation scheme in [6]. We compute the optimal frame and fragment sizes which render the maximum throughput in this context, and compare the results to the optimal values from [6], where AES encryption is not used. We derive asymptotic results related to the MAC layer efficiency, expected frame size and saturation throughput.II. RELATED WORK With respect to increasing efficiency at the MAC layer, much of the previous work has focused on minimizing the contention time which contributes to the transmission overhead ([3], [8], [16], 19). However, Xiao et al. in [17,18,14] show that a theoretical throughput upper limit exists, indicating that by simply increasing the data rate without reducing overhead, the enhanced performance, in terms of throughput and delay, is bounded even when the data rate goes

Energy consumptions analysis for a class of symmetric encryption algorithm

“…The complexity analysis for some of the algorithms under study had attracted many researchers, who developed and used different methodologies to analyze and compute the number of instructions required to encrypt a certain amount of data [8]- [10]. However, those methodologies lacked the comprehensiveness in their analysis, due to the fact that the memory access time had been ignored, which greatly affects the analysis.…”

Trade-off of security and performance of lightweight block ciphers in Industrial Wireless Sensor Networks

“…The papers [36,37] study the optimal network performance for stream ciphers. The paper [38,39] studies the security overhead of aggregation in WIFI. The paper [40] studies the security trade-off of AES over IEEE 802.15.3 wireless personal area networks.…”