Search

The primary requirements of a secure Wireless Sensor Network architecture are confidentiality, integrity and authentication of users and other participating entities. User Authentication for wireless sensor networks is a fundamental and important issue in designing dependable and secure systems. In this thesis, we have outlined the security model, functional requirements, assumptions and network setup for an authentication scheme in the first phase. Keeping in mind the security requirements as well as the flaws of past authentication schemes, we propose a robust user authentication method that inherits user anonymity, mutual authentication and password changing functionality of previous password-based schemes and improves security by resisting gateway bypass and replay attack, and many logged in user with the same ID threat. Our scheme is a variant of strong password based schemes that does not require strict network synchronization. In the second phase of the thesis, we have analysed our authentication scheme from the perspective of security issues and functional requirements. The proposed scheme is modelled in SystemC. It is evaluated in different attack scenarios. The authentication latency, memory and functional requirements, and computational overhead are the metrics used to evaluate the scheme. The effect of multiple users on authentication latency in our scheme is also studied. Some of the past representative schemes have also been modelled and evaluated in the same environment. A detailed comparison of over-head cost, authentication latency and security features are provided in this thesis. It is verified and confirmed by modeling that our scheme provides enhanced security without adding extra computation at the sensor node.

The 802.11ac version of the popular IEEE 802.11 protocol aims to boost performance by increasing the channel bandwidth and allowing simultaneous transmission to multiple clients in the downlink direction. This dissertation presents an innovative approach of performance evaluation of downlink multi-user multiple input multiple output (DL-MUMIMO) technique for differentiated quality of service (QoS) based traffic categories in IEEE 802.11ac protocol. We propose a novel analytical model based on discrete Markov chain (DMC) and E-limited M/G/1 queuing model to evaluate the performance improvements of multi-user transmission opportunity (MU-TXOP) sharing in DL-MU-MIMO under non-saturated load and non-ideal channel condition. We also evaluate MU-TXOP sharing probabilities among different QoS nodes to assess the performance improvement of different traffic categories. In this dissertation, we propose an Access Point controlled MAC protocol (A-MAC) that enables simultaneous transmissions from multiple stations (STA) in uplink to eliminate under utilization of network resources in uplink transmission due to single user communication. The protocol uses enhanced distributed channel access (EDCA) technique to initiate multi-user transmission and orthogonal frequency division multiple access (OFDMA) method to transmit multiple Ready-To-Send (RTS) messages simultaneously. The proposed protocol also introduces explicit channel sounding technique by using dedicated OFDM subcarrier blocks for each user. We evaluate the performance of the proposed protocol using a noble analytical model and validate the performance metrics by extensive simulation in different traffic conditions. In this dissertation we also propose the increase of carrier sensing threshold (CSTH) of nodes during association with access point (AP) to mitigate collision probability due to hidden nodes during uplink transmission and validate our proposal through extensive simulation. Finally, we propose a noble analytical model to evaluate the performance of restricted access window (RAW) mechanism of IEEE 802.11ah as the MAC layer protocol for internet of things (IoT) network. We evaluate detailed performance metrics of non-QoS IoT network and investigate the feasibility of using RAW mechanism to support differentiated QoS based heterogeneous IoT network

The IEEE 802.11ac amendment enhances WLAN throughput by exploiting the spatial diversity of the antennas in a multi-user multi-input multi-output downlink transmission. Still, network resources remain under-utilized in uplink transmission due to single-user communication. In this paper, we propose an access point-controlled MAC protocol (A-MAC) that enables simultaneous transmissions from multiple STAs in uplink. The protocol uses the EDCA channel access technique to initiate multi-user transmission and the OFDMA method to transmit multiple RTSs simultaneously. It also introduces the explicit channel sounding technique by using dedicated OFDM subcarrier blocks for each user. Performance measurement shows that network throughput of the A-MAC is 150% higher than that of a single uplink transmission, thanks to the availability of concurrent multiple RTS transmissions in the uplink. The proposed protocol shortens the backoff time by up to 50% for all traffic categories due to concurrent multiple transmissions and thus enhances the battery life of the nodes. We observe that although the smaller backoff window of high-priority traffic category enhances the network throughput, higher intensity of high-priority traffic drives the network faster to saturation. Furthermore, better network stability and fairness among different traffic categories can be achieved when the dominant traffic has low priority.

The constrained application protocol (CoAP) observe feature has the potential to achieve fast communication and low energy consumption. We have analyzed a CoAP proxy which stores observed data from multiple Internet of Things (IoT) nodes that belong to single IoT domain. To maintain the freshness of cached records under random data arrivals, we have implemented the evaluation of Max-Age parameter at the proxy. We have also implemented congestion control at both IoT nodes and proxy, and show that it interacts with data freshness estimation at the proxy. Our results demonstrate the capacity limits of a single IoT domain against cluster size, interobservation time, proxy re-registration rate, ratio of confirmable/nonconfirmable observations, and presence of congestion control.

In this work we provide comprehensive analytical model for Bitcoin distribution network. We apply Jackson network model on the whole Bitcoin network where individual nodes operate as priority M/G/1 queuing systems. Data arrival process to the nodes is modeled as a non-homogeneous Poisson process in which the data arrival rates to the nodes are derived from the analytical model of gossip data delivery protocol. This model considers random probability distribution of node connectivity. Performance results include network distribution time for blocks, node response time for blocks, and populations of data distribution algorithms as functions on network size. Usefulness of this model is demonstrated by efficiently computing the forking probability for the Bitcoin blockchain.

IEEE 802.11ah protocol aims at providing network connectivity to a large number of end points while maintaining a satisfactory network performance. The RAW scheme of MAC access technique enables STAs to contend in different RAW slots and thereby reducing the chance of collision. In this paper we evaluate the performance metrics of the RAW scheme and observe that RAW scheme may be suitable for very small traffic condition. At higher traffic load, some of the nodes need to wait longer to transmit packets and network becomes unstable very quickly as the traffic for each node is increased beyond two packets per second in our test condition. The highest normalized network throughput achieved in this network is around 40%. Some important performance metrics like idle probability, backoff time and transmission probability are evaluated and studied in detail in this paper.

In this work we discuss architectural alternatives for the design of a proxy that interconnects IoT domains running CoAP with the rest of the Internet including micro datacenters and other domains building scalable hierarchical architectures. We assume that the CoAP domain is terminated by an IoT proxy with cache, and we investigate several design alternatives, assuming that the proxy autonomously maintains data freshness. Our analysis indicates that multicast and observe-based proxies perform better than the default POST/GET approach in terms of successful data transmission, round trip delay and energy consumption, with the multicast option having a slight advantage.

IEEE 802.11ac amendment enhances WLAN throughput by exploiting spatial diversity of antennas in MU-MIMO downlink transmission. However, in the existing framework ample network resources remain underutilized in uplink transmission due to single user communication. In this paper, we have proposed an efficient Access Point controlled MAC (A-MAC) protocol that enables concurrent multiple uplink transmissions for different priority traffic classes. Performance measurement shows that network throughput of A-MAC is 150\% higher than the throughput of single uplink transmission in existing IEEE 802.11ac protocol. We have achieve the highest payload throughput of 64\% when the number of nodes $N_{sum}=8$ and $\lambda=138$ packet/s at uniformly varying packet arrival rates for all STAs. As the number of nodes increases, the network throughput gradually decreases but the packet handling capacity of the network increases. The peak network capacity of 1200 packets/s is achieved with $N_{sum}=16$ and $\lambda=75$ packets/s for all STAs. Further increase of the number of nodes results in decreasing both network throughput and packet handling capacity of the network. We have also evaluated the performance metrics of our proposed A-MAC protocol under non-uniformly varying packet arrival rates for a fixed number of nodes $N_{sum}=24$. The analysis shows that when the network has higher intensity of low priority traffic, the aggregate packet handling capacity of the network is 1400 packets/s. However, as the intensity of high priority traffic increases, the aggregate packet handling capacity of the network decreases to 1000 packets/s. The results show that the more the higher priority packets arrive, the faster the network becomes unstable.

In this paper we model and evaluate power ramping as a mechanism for achieving explicit access priority among traffic classes in LTE#x002F;LTE-A, taking into account preamble SINR considerations, preamble collisions, capture effect, and Physical Downlink Control Channel (PDCCH) deficiency. The results show that explicit power ramping can provide reasonable performance differentiation among traffic classes but it still cannot overcome coupling among traffic classes which exists due to preamble collisions and insufficient PDCCH resources.