Earticle

In global open economy, the study of currency exchange rates prediction with acceptable accuracy under floating exchange rates environment becomes an important issue. Exchange rates can affect a large number of economic decision-makings and participants’ behaviors. Due to the rapid dynamic data changes and increasing large amount of data, accurate and effective currency exchange rates prediction is a rather challenging task. In this paper, we proposed a novel cloud computing approach to do linear regression prediction for dynamic currency exchange rates. We adopt an Intelligent Exchange Rates Prediction System (IERPS) based on cloud computing to collect real-time exchange rates information and predict the future exchange rates in efficient computing time. The system can process large amounts of historical and dynamic data more efficiently and accurately. The experimental results showed that the average error ratios of using linear regression are 94.6%, which is a very good performance.

Malware is huge and growing at an exponential pace. Symantec observes 403 million new malware samples in 2011. Therefore, that efficiently and effectively analysis so many malware samples becomes a great challenge. Centralized systems cause problems of single point of failure as well as processing bottlenecks. Previous distributed systems are mainly applied for specific or simple malware. This paper presents SCMA, a new distributed malware analysis system which can analyze various malware, shares behavior fragments among its monitors efficiently, analyzes malware based on global behavior of malware and aggregates those analyses among monitors in a load-balance way. We implemented a proof-of-concept version of SCMA and performed experiments with 917 real-world malware samples; preliminary results from our evaluation confirm that SCMA has comparable performance with centralized system, but much better scalability, and is approximately consistent with the analysis of AV scanners.

Computing Grid is a high performance computing environment that allows sharing of geographically distributed resources across multiple administrative domains and used to solve large scale computational demands. To achieve the promising potentials of computational grids, job scheduling is an important issue to be considered. This paper addresses scheduling problem of independent tasks on computational grids. A Hybrid Discrete Particle Swarm Optimization algorithm (HDPSO) and Min-min algorithm is presented to reduce overall execution time of task.

Peer-to-peer (P2P) File delivery systems of distributed computing are designed with the understanding that any peer can leave the network at any time, often right after completing its download. The conventional approach to P2P scheduling, shortest processing-time first, is not well suited to peer-leaving situations. We therefore propose a new scheduling algorithm that significantly reduces average finish time in such cases. We further note, at the end of the paper, that our method is also suitable for the more general problem of a dynamic network of peers that may leave early or enter late.

To meet the demands of the large-scale network simulation platform technology, we have improved the classical Min-min algorithm widely used in Grid application. We proposed a novel multi-task network simulation scheduling algorithm based on multi-valued mapping and the new algorithm is called MUNS-Min-min, we also discussed how to determine the weight of simulation time and the resource consumption. Experimental results show that our algorithm is suitable for complex computing environment, and the performance of the simulation platform has improved nearly 20%, compared with traditional scheduling algorithms.

As IP technologies have provided the ability to establish dynamic security associations between endpoints emerge, virtual private networks (VPNs) are going through significant growth. The number of endpoints per virtual network is growing, and the resource management is becoming increasingly hard to predict. To study these issues, we abstract the distributed endpoints into a simple undirected graph, research on the graph and propose an adaptive searching method for resource discovery and management. The new searching strategy reduces the traffic impact to network and improves the ergodic theorem. We also proposed a new generation method, which could be used before traversed the resource. The simulation shows that the result is better than the known.

Failures are inevitable in wireless sensor networks due to the fragile features and unattended deployment. To explore an effectual technique for node’s fault detection caused by energy consumption, a new algorithm, Fault Detection Technique based on Numerical Taxonomy (FDNT) is proposed in this paper. The algorithm is deployed on the sink node with unconstrained energy consumption and is on the basis of dividing WSNs into clusters according to their geographical distribution. Simulation results indicate that this algorithm can get high detection accuracy and will consume less energy.

MapReduce has become an important distributed processing model for large-scale data-intensive applications like data mining and web indexing. Hadoop–an open-source implementation of MapReduce is widely used for short jobs requiring low response time. In this paper, we proposed a new preshuffling strategy in Hadoop to reduce high network loads imposed by shuffle-intensive applications. Designing new shuffling strategies is very appealing for Hadoop clusters where network interconnects are performance bottleneck when the clusters are shared among a large number of applications. The network interconnects are likely to become scarce resource when many shuffle-intensive applications are sharing a Hadoop cluster. We implemented the push model along with the preshuffling scheme in the Hadoop system, where the 2-stage pipeline was incorporated with the preshuffling scheme. We implemented the push model and a pipeline along with the preshuffling scheme in the Hadoop system. Using two Hadoop benchmarks running on the 10-node cluster, we conducted experiments to show that preshuffling-enabled Hadoop clusters are faster than native Hadoop clusters. For example, the push model and the preshuffling scheme powered by the 2-stage pipeline can shorten the execution times of the WordCount and Sort Hadoop applications by an average of 10% and 14%, respectively.

Use of mobile devices for data collection in wireless sensor networks (WSNs) has recently drawn much attention. Controlled sink mobility has been shown to be beneficial in energy conservation and lifetime prolongation. In this paper, we aim at studying the multi-hop uneven hierarchical routing protocol (MUHRP) using different mobile sink strategies. We first analyze the performance of multi-hop uneven hierarchical routing protocol with a fixed sink node. Then to measure the network performance, we use a mobile sink node, replacing the fixed sink node, to collect fused data packets from cluster heads. Network performance using our proposed protocol is validated through simulation experiments using MATLAB.

This paper explores basic level of semantic structure formation in the human vision inferential processes in line with Gestalt laws and proposes micro level semantic structure formations and their relational combinations. Using this approach two sets of semantic features have been derived for visual object class recognition. The first algorithm uses the hypothesis in line with Gestalt laws of proximity that; in an image, basic semantic structures are formed by line segments (arcs also approximated and broken into smaller line segments based on pixel deviation threshold) which are in close proximity of each other. Based on the notion of proximity a transitive relation is defined, which combines basic micro level semantic structures hierarchically till such a point where semantic meanings of the structure can be extracted. The algorithm extracts line segments in an image and then forms semantic groups of these line segments based on a minimum distance threshold from each other. The line segment groups so formed can be differentiated from each other, by the number of group members and their geometrical properties. The geometrical properties of these semantic groups are used to generate rotation, translation and scale invariant histograms used as feature vectors for object class recognition tasks in a K-nearest neighbor framework. In the second approach a semantic group based on the proximity distance is clustered and modeled as a graph vertex. The line segments which are common to more than one semantic group are defined as semantic relations between the semantic groups and are modeled as edges of the graph. This way an image object is transformed into a graph using micro level structure formations. Each vertex and edge is labeled using translation, rotation and scale invariant properties of the member segments of each vertex and edge. From a set of training images, a graph model is constructed for visual object class recognition. The graph model is constructed by iteratively combining the training graphs and frequency labeling the vertices and edges. After the combining phase, all the vertices and edges whose repetition frequency is below a threshold are removed. The final graph model consists of the semantic nodes which are highly common in the training images. The recognition is based on graph matching the query image graph and the model graph. The model graph generates a vote for the query and ties are resolved by considering the node frequencies in the query and model graph. The algorithms have been applied to classify 101 object classes at one time. The results have been compared with existing state of the art approaches and are found promising. Results from above approaches show that low level image structure and other features can be used to construct different type of semantic features, which can help a model or a classifier make more intelligent decisions and work more effectively for the task compared to low level features alone. Our experimental results are comparable, or outperform other state-of-the-art approaches. We have also summarized the state-of-the-art at the time this work was finished. We conclude with a discussion about the possible future extensions.

Recently, user-oriented QoS requirements have attracted more and more attentions, and the resource cost has become the key QoS attribute in many practical grid systems. However, how to decide the resource prices when co-allocating plenty of heterogeneous resources across different virtual organizations remains a challenging issue. In this work, we design and implement a novel co-allocation framework, in which the resource co-allocation procedure is divided into two phases. In the first phase, resource providers uses co-operative gaming model to decide their resource’s original price, which can lead to maximal benefits for resource providers; In the second phase, non-cooperative gaming model is applied to find out the retail price (trading price) when users buy resources for executing their applications. Extensive simulations are conducted to evaluate the effectiveness and performance of the proposed framework, and the results show that the two-phase model can significantly improve the QoS satisfaction for those grid applications with constraint to limited budgets. Also, the efficiency of co-allocating multiple resources is also significantly improved comparing with existing policies.