Many of the problems arising from rapid urbanization and urban population growth can be solved by making cities “smart”. These smart cities are supported by large networks of interconnected and widely geo-distributed devices, known as Internet of Things or IoT, that generate large volumes of data. Traditionally, cloud computing has been the technology used to support this infrastructure; however, some of the essential requirements of smart cities such as low-latency, mobility support, location-awareness, bandwidth cost savings, and geo-distributed nature of such IoT systems cannot be met. To solve these problems, the fog computing paradigm proposes extending cloud computing models to the edge of the network. However, most of the proposed architectures and frameworks are based on their own private data models and interfaces, which severely reduce the openness and interoperability of these solutions. To address this problem, we propose a standard-based fog computing architecture to enable it to be an open and interoperable solution. The proposed architecture moves the stream processing tasks to the edge of the network through the use of lightweight context brokers and Complex Event Processing (CEP) to reduce latency. Moreover, to communicate the different smart cities domains we propose a Context Broker based on a publish/subscribe middleware specially designed to be elastic and low-latency and exploit the context information of these environments. Additionally, we validate our architecture through a real smart city use case, showing how the proposed architecture can successfully meet the smart cities requirements by taking advantage of the fog computing approach. Finally, we also analyze the performance of the proposed Context Broker based on microbenchmarking results for latency, throughput, and scalability.

No abstract

No abstract

Outlier detection is an important problem occurring in a wide range of areas. Outliers are the outcome of fraudulent behaviour, mechanical faults, human error, or simply natural deviations. Many data mining applications perform outlier detection, often as a preliminary step in order to filter out outliers and build more representative models. In this paper, we propose an outlier detection method based on a clustering process. The aim behind the proposal outlined in this paper is to overcome the specificity of many existing outlier detection techniques that fail to take into account the inherent dispersion of domain objects. The outlier detection method is based on four criteria designed to represent how human beings (experts in each domain) visually identify outliers within a set of objects after analysing the clusters. This has an advantage over other clustering-based outlier detection techniques that are founded on a purely numerical analysis of clusters. Our proposal has been evaluated, with satisfactory results, on data (particularly time series) from two different domains: stabilometry, a branch of medicine studying balance-related functions in human beings and electroencephalography (EEG), a neurological exploration used to diagnose nervous system disorders. To validate the proposed method, we studied method outlier detection and efficiency in terms of runtime. The results of regression analyses confirm that our proposal is useful for detecting outlier data in different domains, with a false positive rate of less than 2% and a reliability greater than 99%.

No abstract