Machine learning and deep learning algorithms can be used to classify encrypted Internet traffic. Classification of encrypted traffic can become more challenging in the presence of adversarial attacks that target the learning algorithms. In this paper, we focus on investigating the effectiveness of different evasion attacks and see how resilient machine and deep learning algorithms are. Namely, we test C4.5 Decision Tree, K-Nearest Neighbor (KNN), Artificial Neural Network (ANN), Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN). In most of our experimental results, deep learning shows better resilience against the adversarial samples in comparison to machine learning. Whereas, the impact of the attack varies depending on the type of attack.

Malicious co-residency in virtualized networks poses a real threat. The next-generation mobile networks heavily rely on virtualized infrastructure, and network slicing has emerged as a key enabler to support different virtualized services and applications in the 5G network. However, allocating network slices efficiently while providing a minimum guaranteed level of service, as well as providing defense against the threat of malicious co-residency in a mobile core network, is challenging. To address these questions, in our previous work, we proposed an optimization model to allocate slices. It provided a static and manual allocation of slices. In this work, we analyze the defense against the malicious co-residency using our optimization-based allocation, and we extend our work to dynamically allocate slices. We propose a dynamic slice allocation framework for the 5G core network. The proposed framework provides user-interaction to request slices and any required services that need to run on a slice(s). It can accept a single or multiple allocation requests, and it dynamically allocates them. Additionally, the framework allocates slices in a balanced fashion across available resources. We compare our framework with the First Come First Serve and First Available allocation scheme.

Network virtualization offers flexibility by decoupling virtual network from the underlying physical network. Software-Defined Network (SDN) could utilize the virtual network. For example, in Software-Defined Networks, the entire network can be run on commodity hardware and operating systems that use virtual elements. However, this could present new challenges of data plane performance. In this paper, we present an empirical model of the packet processing delay of a widely used OpenFlow virtual switch, the Open vSwitch. In the empirical model, we analyze the effect of varying Random Access Memory (RAM) and network parameters on the performance of the Open vSwitch. Our empirical model captures the non-network processing delays, which could be used in enhancing the network modeling and simulation.

Time synchronization plays an important role in distributed systems. Distributed wireless sensor networks (WSN) often require accurate time synchronization for coordination and data reliability. But precision of time synchronization is limited to scope and criticality of the application. Energy, lifetime and time synchronization are important parameters of any wireless sensor network. Flooding Time Synchronization Protocol (FTSP) and Recursive Time Synchronization Protocol (RTSP) are two state of the art protocols for time synchronization in WSNs. In this paper, a comparative analysis of both protocols under two different topologies i.e. bus and tree, using message count as performance parameter. Xbee Pro and Arduino Mega are used for communication and logical design implementation of both protocols, respectively.