Transactions on Computational Collective Intelligence XXXIII

In this paper, we propose a novel framework for performance optimization in Internet of Things (IoT)-based next-generation wireless sensor networks. In particular, a computationally-convenient system is presented to combat two major research problems in sensor networks. First is the conventionally-tackled resource optimization problem which triggers the drainage of battery at a faster rate within a network. Such drainage promotes inefficient resource usage thereby causing sudden death of the network. The second main bottleneck for such networks is the data degradation. This is because the nodes in such networks communicate via a wireless channel, where the inevitable presence of noise corrupts the data making it unsuitable for practical applications. Therefore, we present a layer-adaptive method via 3-tier communication mechanism to ensure the efficient use of resources. This is supported with a mathematical coverage model that deals with the formation of coverage holes. We also present a transform-domain based robust algorithm to effectively remove the unwanted components from the data. Our proposed framework offers a handy algorithm that enjoys desirable complexity for real-time applications as shown by the extensive simulation results.

Service oriented environments such as cloud computing infrastructures aim at facilitating the requirements of users and enterprises by providing services following an on-demand orientation. While the advantages of such environments are clear and lead to wide adoption, the key concern of the non-adopters refers to privacy and security. Even though providers put in place several measures to minimize security and privacy vulnerabilities, the users are still in many cases reluctant to move their data and applications to clouds. In this paper an approach is presented that proposes the use of security controls as plugins that can be ingested in service-oriented environments. The latter allows users to tailor the corresponding security and privacy levels by utilizing security measures that have been selected and implemented by themselves, thus alleviating their security and privacy concerns. The challenges and an architecture with the corresponding key building blocks that address these challenges are presented. Furthermore, results in the context of trustworthy requirements, i.e. dependability, are presented to evaluate the proposed approach.

We consider the P2PPS (peer-to-peer type of topic-based publish/subscribe) model where each peer process (peer) can publish and subscribe event messages with no centralized coordinator. Here, hidden topics are topics which a source peer is allowed to subscribe but a target peer is not allowed to subscribe. After receipt of an event message \(e_1\) with hidden topics, if a peer publishes another event message \(e_2\), the event message \(e_2\) may be related with the hidden topics of the event message \(e_1\). Hence, if an event message with hidden topics is received by another target peer which does not subscribe the hidden topics, the target peer can get information on the hidden topics. This means, illegal information flow to the target peer occurs. However, some hidden topics may be related with a subscription topic of a target peer and the target peer just may not know about the hidden topics. In this paper, we newly introduce a learning mechanism where each peer newly obtains hidden topics if the hidden topics are related with subscription topics. In this paper, we newly propose an FS-H (flexible synchronization for hidden topics) protocol. In the evaluation, we show the fewest number of event messages are prohibited from being received in the FS-H protocol compared with the other protocols.

In cloud computing domain, often service providers offer services with same functionalities, but with varying quality metrics. A suitable service selection method finds the most appropriate solution among the alternatives. The challenge is to deliver a solution satisfying the requirement (quality and other) of a consumer with minimum possible execution time. Many conflicting QoS objectives increase the complexity of the problem. In fact, the problem may be formulated as a multi-objective, NP-hard optimization problem. Most of the existing solutions either satisfies the QoS demands of consumer or only reduces execution time by considering a sub-set of required QoS metrics. Consumer’s feedback on the choice of required QoS metrics not only shall help increasing user satisfaction, but also may reduce the complexity effectively. However, this depends on the domain knowledge of a consumer. In this work, we have proposed a goodness measure that replaces all QoS metrics by a single one. The new technique using dimension reduction is proposed to offer significant improvement compared to the existing works in terms of execution time. Moreover, the solution satisfies all the QoS requirements of a consumer in most of the cases. The proposed data driven selection approach has been implemented and the experimental results substantiate the claims as mentioned.

This paper shows the effect of the controlling file assignment on the file transfer time for a Web-based content categorization system. Our proposed algorithm estimates categories of contents based on the terms and the content categories already added. However, our algorithm uses a large table that consists of the scores that represent the relationship between a term and a category. To address the large table size and longer calculation time, we proposed a distributed Web system that uses multiple calculation machines. This Web system runs preprocessors on a Web browser and calculation machines. In this Web system, the file transfer time becomes a problem when a user sends larger files. In this paper, we propose a way to resolve the issue of longer file transfer time by controlling the file assignment. We assign the large files to the Web browser process, and we assign the smaller files to the calculation machines over the network.

The web is a complex information ecosystem that provides a large variety of content changing over time as a consequence of the combined effects of management policies, user interactions and external events. These highly dynamic scenarios challenge technologies dealing with discovery, management and retrieval of web content. In this paper, we address the problem of modeling and predicting web dynamics in the framework of time series analysis and forecasting. We present a general methodological approach that allows the identification of the patterns describing the behavior of the time series, the formulation of suitable models and the use of these models for predicting the future behavior. Moreover, to improve the forecasts, we propose a method for detecting and modeling the spiky patterns that might be present in a time series. To test our methodological approach, we analyze the temporal patterns of page uploads of the Reuters news agency website over one year. We discover that the upload process is characterized by a diurnal behavior and by a much larger number of uploads during weekdays with respect to weekend days. Moreover, we identify several sudden spikes and a daily periodicity. The overall model of the upload process – obtained as a superposition of the models of its individual components – accurately fits the data, including most of the spikes.

In prevent global warming, it is critical to reduce electric energy consumed in information systems, especially servers in clusters like cloud computing systems. In this paper, a process migration approach is discussed to reduce the total energy consumption of clusters by using virtual machines. We propose a pair of the static SM(v) and dynamic DM(v) migration algorithms where a group of at most v (\(\ge \)0) virtual machines migrate from a host server to a guest server. A group of virtual machines on a host server to migrate to a guest server are selected so that the total energy to be consumed by the host and guest servers can be reduced. In the SM(v) algorithm, the total number of virtual machines is fixed in a cluster. In the DM(v) algorithm, virtual machines are resumed and suspended so that the number of processes on each virtual machine is kept fewer. In the evaluation, we show the total energy consumption of servers can be mostly reduced in the DM(v) algorithm compared with other algorithms.

Software-Defined Networks (SDN) initiates a novel networking model. SDN introduces the separation of forwarding and control planes by proposing a new independent plane called network controller. The architecture enhances the network resilient, decompose management complexity, and support more straightforward network policies enforcement. However, the model suffers severe security threats. Specifically, a centralized network controller is a precious target for the attackers for two reasons. First, the controller is located at a central location between the application and data planes. Second, a controller is software which prone to vulnerabilities, e.g., buffer and stack overflow. Hence, providing security measures is a crucial procedure towards the fully unleash of the new model capabilities. Intrusion detection is one option to enhance networking security. Several approaches were proposed, for instance, signature-based, and anomaly detection. Anomaly detection is a broad approach deployed by various methods, e.g., machine learning. For many decades intrusion detection solution suffers performance and accuracy deficiencies. This paper revisits network anomalies detection as recent advances in machine learning particularly deep learning. The study proposes an intrusion detection framework based on unsupervised deep learning algorithms. The framework consists of an unsupervised deep learning phase followed by simple clustering algorithms, e.g. k-means. Our results showed accuracy over 99%, that is a significant improvement in detection accuracy.