Fan-Hsun Tseng

A survey of black hole attacks in wireless mobile ad hoc networks

The black hole attack is one of the well-known security threats in wireless mobile ad hoc networks. The intruders utilize the loophole to carry out their malicious behaviors because the route discovery process is necessary and inevitable. Many researchers have conducted different detection techniques to propose different types of detection schemes. In this paper, we survey the existing solutions and discuss the state-of-the-art routing methods. We not only classify these proposals into single black hole attack and collaborative black hole attack but also analyze the categories of these solutions and provide a comparison table. We expect to furnish more researchers with a detailed work in anticipation.

Ultra-dense small cell planning using cognitive radio network toward 5G

Mobile communication is facing new challenges to the soaring traffic demand of numerous user devices; thus, the notion of the small cell has been proposed and realized in recent years. However, licensed spectrum has been occupied by various underlying access technologies, so the deployment of small cells needs a sophisticated planning algorithm. In this article, we provide an overview of reconfigurable radio and small cell technologies, then introduce the tentative network architecture for 5G. Two planning approaches (i.e., genetic-based and graphbased) are proposed that accommodate cognitive radio technology to improve user throughput by eliminating communication interference. Since cognitive radio networking provides frequency allocation with cognition cycle for better spectral efficiency, we tackle the deployment of ultradense small cells and consider the coordination of unlicensed spectrum at the same time. Results show that the proposed algorithms with spectrum cognition improve network performance in terms of throughput and signal-to-interference- plus-noise ratio. Specifically, the genetic- based algorithm increases 232 percent in throughput and 150 percent in signal-to-interference- plus-noise ratio compared to the graphbased algorithm. Finally, we conclude this article by discussing potential challenges and opportunities.

A novel VM workload prediction using Grey Forecasting model in cloud data center

In recent years, the resource demands in cloud environment have been increased incrementally. In order to effectively allocate the resources, the workload prediction of virtual machines (VMs) is a vital issue that makes the VM allocation more instantaneous and reduces the power consumption. In this paper, we propose a workload prediction method using Grey Forecasting model to allocate VMs, which is the first string in the research field. Firstly, we utilize the time-dependent of workload at the same period in every day, and forecast the VM workload tendency towards increasing or decreasing. Next, we compare the predicted value with previous time period on workload usage, then determine to migrate which VM wherein the physical machine (PM) for the balanced workload and lower power consumption. The simulation results show that our proposed method not only uses the fewer data to predict the workload accurately but also allocates the resource of VMs with power saving.

Efficient Power Conservation Mechanism in Spline Function Defined WSN Terrain

Coverage is a vital issue that ensures that basic functions are available in wireless sensor networks (WSNs). These functions provide communications during emergency rescue or in war environments. Sensor nodes must be dispersed and survive in place for a long time in order to accurately monitor all events. However, sensor nodes rely on limited battery energy restricted to the lifetime of the entire network. When few sensor nodes are awake during an epoch, the entire network lifetime becomes longer. This implies that the coverage problem must be solved based on the energy efficiency issue. In order to strengthen the coverage ratio with a limited number of sensor nodes, practical coverage algorithms were proposed in the existing researches but investigated in a 2-D plane, which is not suitable for a realistic environment. Few researchers have investigated the WSN coverage problem for a 3-D space. The actual terrain is rugged and bumpy that prevents deploying sensor nodes in three-dimension space. In this paper, we consider the field of interest as a complex surface but also use the spline function to obtain factual information about a complex surface. Furthermore, dynamic programming is utilized to optimize the sensing radius for reducing the power consumption.

MR-Chord: Improved Chord Lookup Performance in Structured Mobile P2P Networks

Peer-to-peer (P2P) networks are becoming very popular since various applications such as media streaming and voice over IP use these networks in different environment settings without the need for a client-server structure. P2P protocols have been originally designed for traditional wired networks, and when deployed in wireless network environments, several challenges are encountered. For instance, P2P clients may depart or join the network frequently, raising the issue of identification and retrieval of data items in an efficient manner. In this scenario, the routing information in P2P clients may become overdue, leading to lookup failures. This paper continues the investigation of our recently proposed solution for Chord lookup in mobile P2P networks [so-called mobile robust Chord (MR-Chord)]. MR-Chord was designed to maintain and update the finger table using a modified distributed hash table-based protocol, so that the necessary lookup services in the network are provided. Our contribution consists in studying the effects of node mobility on the performance of MR-Chord. Simulation results show that in the presence of node mobility, MR-Chord outperforms the original Chord protocol in terms of lookup success rate, overlay consistency, lookup delay time, lookup hot count, and total network load, chosen as performance metrics.

Implement a reliable and secure cloud distributed file system

In recent years, there are various cloud storage services have been proposed. The users can acquire the enormous cloud storage spaces through Internet even if the storage space of portable devices and handsets are limited. In this paper, we implement a distributed file system based on Apache Hadoop, and the original Hadoop distributed file system is replaced with the Tahoe least-authority file system. Furthermore, we investigate the total transmission rate and download time with two different file size. The experiment result shows that the established file system provides a secure and reliable cloud storage service even if parts of storage nodes are broken.

A Study on Coverage Problem of Network Planning in LTE-Advanced Relay Networks

In recent years new research studies have appeared that concern the issue of network planning in LTE-Advanced. In this paper, the coverage problem in LTE-Advanced relay networks is formulated based on integer linear programming (ILP). We propose the Enhanced tree (E-Tree) algorithm to place the evolved Node B (eNB) and relay station (RS) at the location which has the lowest construction cost. The goal is not only satisfy the minimum requirement but also meet the two hop relaying limitation in LTE-Advanced networks. The E-Tree algorithm is proposed based on graph theoretic technique, and analyzed with the simulation results. The simulation results show that the proposed algorithm provides a rapid planning method and the lowest construction cost with various network environment.

Service-Oriented Virtual Machine Placement Optimization for Green Data Center

The first service-oriented virtual machine (VM) placement for green data center is designed in this work. Integer Linear Programming (ILP) is the problem design basis. The Tree algorithm is proposed to place VM role instances at the lowest communication cost, economizing the construction cost with fewer physical servers. Another Forest algorithm is also proposed for balancing the computation load between the physical machines. Both of the proposed algorithms are formulated on the graph theoretic technique and evaluated and analyzed using the Best Fit algorithm in the simulations. Although the total power consumption and average utility of both proposed algorithms are slightly impaired, the unnecessary outbound communication cost is significantly eliminated and decreased, especially in the immense number of services. The results show that the proposed Tree and Forest algorithms provide lower communication cost than the Best Fit algorithm, and achieve green communications for large scale environment such as cloud or green data center.

Minimizing Virtual Machine Migration Probability for Cloud Environments

In recent years, many researchers focus on how to balance the number of average virtual machines (VMs) in cloud data center. However, the VMs suddenly jumped to more than the utilization threshold of the physical machine (PM) and exhausted the computing resource, which will lead to VM migration and extra cost. In this paper, the VM migration problem in cloud computing is formulated based on mixed integer linear programming (MILP). Then we propose an Allocation algorithm to distribute the VMs among the PMs, which is based on the Support Vector Machine (SVM). The ultimate goal that not only minimizes the VM migration probability but also maximizes the PM resources as much as possible. According to the training data in a period time, the classified VMs can be allocated into the PMs with the lowest VM migration probability by the proposed Allocation algorithm. The simulation results show that the proposed algorithm can achieve lower VM migration probability and reduce the cost for the cloud environments.

A Novel Method Defends against the Path-Based DoS for Wireless Sensor Network

WSNs are significantly different from the traditional network architecture due to its wireless communication, energy limitation, and computation constraint and environment of the application. Because of these differences, security becomes a critical issue. The path-based denial of service (PDoS) attacks harm the network maintenance and cause serious damage in the resource constrained WSNs. In a PDoS attack, an adversary can overwhelm sensor node and cluster head node to flood packets along the routing path so that intermediate node must keep active mode and exhaust the energy. In this paper, we creatively propose a novel method, which is operated at the base station to detect the malicious behavior. The proposed method is combined with triple exponential smoothing and Markov chain, so that it makes the detection results more accurate. Meanwhile, we first use the concept of black hole to defend the PDoS attack in WSNs. Simulation results are provided to evaluate the performance and illustrate the contribution of this mechanism.