Building a microscope for the data center

Managing the physical and compute infrastructure of a large data center is an embodiment of a Cyber-Physical System (CPS). The physical parameters of the data center (such as power, temperature, pressure, humidity) are tightly coupled with computations, even more so in upcoming data centers, where the location of workloads can vary substantially due, for example, to workloads being moved in a cloud infrastructure hosted in the data center. In this paper, we describe a data collection and distribution architecture that enables gathering physical parameters of a large data center at a very high temporal and spatial resolutionof the sensor measurements. We think this is an important characteristic to enable more accurate heat-flow models of the data center andwith them, _and opportunities to optimize energy consumption. Havinga high resolution picture of the data center conditions, also enables minimizing local hotspots, perform more accurate predictive maintenance (pending failures in cooling and other infrastructure equipment can be more promptly detected) and more accurate billing. We detail this architecture and define the structure of the underlying messaging system that is used to collect and distribute the data. Finally, we show the results of a preliminary study of a typical data center radio environment.

Scalable data acquisition for densely instrumented cyber-physical systems

Consider the problem of designing an algorithm for acquiring sensor readings. Consider specifically the problem of obtaining an approximate representation of sensor readings where (i) sensor readings originate from different sensor nodes, (ii) the number of sensor nodes is very large, (iii) all sensor nodes are deployed in a small area (dense network) and (iv) all sensor nodes communicate over a communication medium where at most one node can transmit at a time (a single broadcast domain). We present an efficient algorithm for this problem, and our novel algorithm has two desired properties: (i) it obtains an interpolation based on all sensor readings and (ii) it is scalable, that is, its time-complexity is independent of the number of sensor nodes. Achieving these two properties is possible thanks to the close interlinking of the information processing algorithm, the communication system and a model of the physical world.

Scalable and efficient data processing in networked control systems

Network control systems (NCSs) are spatially distributed systems in which the communication between sensors, actuators and controllers occurs through a shared band-limited digital communication network. However, the use of a shared communication network, in contrast to using several dedicated independent connections, introduces new challenges which are even more acute in large scale and dense networked control systems. In this paper we investigate a recently introduced technique of gathering information from a dense sensor network to be used in networked control applications. Obtaining efficiently an approximate interpolation of the sensed data is exploited as offering a good tradeoff between accuracy in the measurement of the input signals and the delay to the actuation. These are important aspects to take into account for the quality of control. We introduce a variation to the state-of-the-art algorithms which we prove to perform relatively better because it takes into account the changes over time of the input signal within the process of obtaining an approximate interpolation.

Energy efficient scheduling for cluster-tree wireless sensor networks with time-bounded data flows: application to IEEE 802.15.4/ZigBee

Cluster scheduling and collision avoidance are crucial issues in large-scale cluster-tree Wireless Sensor Networks (WSNs). The paper presents a methodology that provides a Time Division Cluster Scheduling (TDCS) mechanism based on the cyclic extension of RCPS/TC (Resource Constrained Project Scheduling with Temporal Constraints) problem for a cluster-tree WSN, assuming bounded communication errors. The objective is to meet all end-to-end deadlines of a predefined set of time-bounded data flows while minimizing the energy consumption of the nodes by setting the TDCS period as long as possible. Sinceeach cluster is active only once during the period, the end-to-end delay of a given flow may span over several periods when there are the flows with opposite direction. The scheduling tool enables system designers to efficiently configure all required parameters of the IEEE 802.15.4/ZigBee beaconenabled cluster-tree WSNs in the network design time. The performance evaluation of thescheduling tool shows that the problems with dozens of nodes can be solved while using optimal solvers.

Simulation study of energy efficient scheduling for IEEE 802.15.4/ZigBee cluster-treeWireless Sensor Networks with time-bounded data flows

The simulation analysis is important approach to developing and evaluating the systems in terms of development time and cost. This paper demonstrates the application of Time Division Cluster Scheduling (TDCS) tool for the configuration of IEEE 802.15.4/ZigBee beaconenabled cluster-tree WSNs using the simulation analysis, as an illustrative example that confirms the practical applicability of the tool. The simulation study analyses how the number of retransmissions impacts the reliability of data transmission, the energy consumption of the nodes and the end-to-end communication delay, based on the simulation model that was implemented in the Opnet Modeler. The configuration parameters of the network are obtained directly from the TDCS tool. The simulation results show that the number of retransmissions impacts the reliability, the energy consumption and the end-to-end delay, in a way that improving the one may degrade the others.

Elements of scalable data processing

Cooperating objects (COs) is a recently coined term used to signify the convergence of classical embedded computer systems, wireless sensor networks and robotics and control. We present essential elements of a reference architecture for scalable data processing for the CO paradigm.

Data mining languages for business intelligence

Doctoral Thesis in Information Systems and Technologies Area of Engineering and Manag ement Information Systems

A lightweight indoor localization model based on motley-keenan and COST

This paper presents a novel approach to WLAN propagation models for use in indoor localization. The major goal of this work is to eliminate the need for in situ data collection to generate the Fingerprinting map, instead, it is generated by using analytical propagation models such as: COST Multi-Wall, COST 231 average wall and Motley- Keenan. As Location Estimation Algorithms kNN (K-Nearest Neighbour) and WkNN (Weighted K-Nearest Neighbour) were used to determine the accuracy of the proposed technique. This work is based on analytical and measurement tools to determine which path loss propagation models are better for location estimation applications, based on Receive Signal Strength Indicator (RSSI).This study presents different proposals for choosing the most appropriate values for the models parameters, like obstacles attenuation and coefficients. Some adjustments to these models, particularly to Motley-Keenan, considering the thickness of walls, are proposed. The best found solution is based on the adjusted Motley-Keenan and COST models that allows to obtain the propagation loss estimation for several environments.Results obtained from two testing scenarios showed the reliability of the adjustments, providing smaller errors in the measured values values in comparison with the predicted values.

RSS and LEA adaptation for indoor location using fingerprinting

Fingerprinting is an indoor location technique, based on wireless networks, where data stored during the offline phase is compared with data collected by the mobile device during the online phase. In most of the real-life scenarios, the mobile node used throughout the offline phase is different from the mobile nodes that will be used during the online phase. This means that there might be very significant differences between the Received Signal Strength values acquired by the mobile node and the ones stored in the Fingerprinting Map. As a consequence, this difference between RSS values might contribute to increase the location estimation error. One possible solution to minimize these differences is to adapt the RSS values, acquired during the online phase, before sending them to the Location Estimation Algorithm. Also the internal parameters of the Location Estimation Algorithms, for example the weights of the Weighted k-Nearest Neighbour, might need to be tuned for every type of terminal. This paper focuses both approaches, using Direct Search optimization methods to adapt the Received Signal Strength and to tune the Location Estimation Algorithm parameters. As a result it was possible to decrease the location estimation error originally obtained without any calibration procedure.

A web-based tool to evaluate the iterative processes of direct search methods

Constrained and unconstrained Nonlinear Optimization Problems often appear in many engineering areas. In some of these cases it is not possible to use derivative based optimization methods because the objective function is not known or it is too complex or the objective function is non-smooth. In these cases derivative based methods cannot be used and Direct Search Methods might be the most suitable optimization methods. An Application Programming Interface (API) including some of these methods was implemented using Java Technology. This API can be accessed either by applications running in the same computer where it is installed or, it can be remotely accessed through a LAN or the Internet, using webservices. From the engineering point of view, the information needed from the API is the solution for the provided problem. On the other hand, from the optimization methods researchers’ point of view, not only the solution for the problem is needed. Also additional information about the iterative process is useful, such as: the number of iterations; the value of the solution at each iteration; the stopping criteria, etc. In this paper are presented the features added to the API to allow users to access to the iterative process data.

Analysis of financial data series using fractional Fourier transform and multidimensional scaling

The goal of this study is the analysis of the dynamical properties of financial data series from worldwide stock market indexes during the period 2000–2009. We analyze, under a regional criterium, ten main indexes at a daily time horizon. The methods and algorithms that have been explored for the description of dynamical phenomena become an effective background in the analysis of economical data. We start by applying the classical concepts of signal analysis, fractional Fourier transform, and methods of fractional calculus. In a second phase we adopt the multidimensional scaling approach. Stock market indexes are examples of complex interacting systems for which a huge amount of data exists. Therefore, these indexes, viewed from a different perspectives, lead to new classification patterns.

Calculation of fractional derivatives of noisy data with genetic algorithms

This paper addresses the calculation of derivatives of fractional order for non-smooth data. The noise is avoided by adopting an optimization formulation using genetic algorithms (GA). Given the flexibility of the evolutionary schemes, a hierarchical GA composed by a series of two GAs, each one with a distinct fitness function, is established.

Agile modeling data warehouse development

Most of the traditional software and database development approaches tend to be serial, not evolutionary and certainly not agile, especially on data-oriented aspects. Most of the more commonly used methodologies are strict, meaning they’re composed by several stages each with very specific associated tasks. A clear example is the Rational Unified Process (RUP), divided into Business Modeling, Requirements, Analysis & Design, Implementation, Testing and Deployment. But what happens when the needs of a well design and structured plan, meet the reality of a small starting company that aims to build an entire user experience solution. Here resource control and time productivity is vital, requirements are in constant change, and so is the product itself. In order to succeed in this environment a highly collaborative and evolutionary development approach is mandatory. The implications of constant changing requirements imply an iterative development process. Project focus is on Data Warehouse development and business modeling. This area is usually a tricky one. Business knowledge is part of the enterprise, how they work, their goals, what is relevant for analyses are internal business processes. Throughout this document it will be explained why Agile Modeling development was chosen. How an iterative and evolutionary methodology, allowed for reasonable planning and documentation while permitting development flexibility, from idea to product. More importantly how it was applied on the development of a Retail Focused Data Warehouse. A productized Data Warehouse built on the knowledge of not one but several client needs. One that aims not just to store usual business areas but create an innovative sets of business metrics by joining them with store environment analysis, converting Business Intelligence into Actionable Business Intelligence.

Bridging the gap between closed and open data. System proposal for the Portuguese Legislation

Esta dissertação apresenta uma proposta de sistema capaz de preencher a lacuna entre documentos legislativos em formato PDF e documentos legislativos em formato aberto. O objetivo principal é mapear o conhecimento presente nesses documentos de maneira a representar essa coleção como informação interligada. O sistema é composto por vários componentes responsáveis pela execução de três fases propostas: extração de dados, organização de conhecimento, acesso à informação. A primeira fase propõe uma abordagem à extração de estrutura, texto e entidades de documentos PDF de maneira a obter a informação desejada, de acordo com a parametrização do utilizador. Esta abordagem usa dois métodos de extração diferentes, de acordo com as duas fases de processamento de documentos – análise de documento e compreensão de documento. O critério utilizado para agrupar objetos de texto é a fonte usada nos objetos de texto de acordo com a sua definição no código de fonte (Content Stream) do PDF. A abordagem está dividida em três partes: análise de documento, compreensão de documento e conjunção. A primeira parte da abordagem trata da extração de segmentos de texto, adotando uma abordagem geométrica. O resultado é uma lista de linhas do texto do documento; a segunda parte trata de agrupar os objetos de texto de acordo com o critério estipulado, produzindo um documento XML com o resultado dessa extração; a terceira e última fase junta os resultados das duas fases anteriores e aplica regras estruturais e lógicas no sentido de obter o documento XML final. A segunda fase propõe uma ontologia no domínio legal capaz de organizar a informação extraída pelo processo de extração da primeira fase. Também é responsável pelo processo de indexação do texto dos documentos. A ontologia proposta apresenta três características: pequena, interoperável e partilhável. A primeira característica está relacionada com o facto da ontologia não estar focada na descrição pormenorizada dos conceitos presentes, propondo uma descrição mais abstrata das entidades presentes; a segunda característica é incorporada devido à necessidade de interoperabilidade com outras ontologias do domínio legal, mas também com as ontologias padrão que são utilizadas geralmente; a terceira característica é definida no sentido de permitir que o conhecimento traduzido, segundo a ontologia proposta, seja independente de vários fatores, tais como o país, a língua ou a jurisdição. A terceira fase corresponde a uma resposta à questão do acesso e reutilização do conhecimento por utilizadores externos ao sistema através do desenvolvimento dum Web Service. Este componente permite o acesso à informação através da disponibilização de um grupo de recursos disponíveis a atores externos que desejem aceder à informação. O Web Service desenvolvido utiliza a arquitetura REST. Uma aplicação móvel Android também foi desenvolvida de maneira a providenciar visualizações dos pedidos de informação. O resultado final é então o desenvolvimento de um sistema capaz de transformar coleções de documentos em formato PDF para coleções em formato aberto de maneira a permitir o acesso e reutilização por outros utilizadores. Este sistema responde diretamente às questões da comunidade de dados abertos e de Governos, que possuem muitas coleções deste tipo, para as quais não existe a capacidade de raciocinar sobre a informação contida, e transformá-la em dados que os cidadãos e os profissionais possam visualizar e utilizar.

Recomendações personalizadas de alunos em sistemas de hipermédia adaptativa educacional usando Data Mining

O aumento de tecnologias disponíveis na Web favoreceu o aparecimento de diversas formas de informação, recursos e serviços. Este aumento aliado à constante necessidade de formação e evolução das pessoas, quer a nível pessoal como profissional, incentivou o desenvolvimento área de sistemas de hipermédia adaptativa educacional - SHAE. Estes sistemas têm a capacidade de adaptar o ensino consoante o modelo do aluno, características pessoais, necessidades, entre outros aspetos. Os SHAE permitiram introduzir mudanças relativamente à forma de ensino, passando do ensino tradicional que se restringia apenas ao uso de livros escolares até à utilização de ferramentas informáticas que através do acesso à internet disponibilizam material didático, privilegiando o ensino individualizado. Os SHAE geram grande volume de dados, informação contida no modelo do aluno e todos os dados relativos ao processo de aprendizagem de cada aluno. Facilmente estes dados são ignorados e não se procede a uma análise cuidada que permita melhorar o conhecimento do comportamento dos alunos durante o processo de ensino, alterando a forma de aprendizagem de acordo com o aluno e favorecendo a melhoria dos resultados obtidos. O objetivo deste trabalho foi selecionar e aplicar algumas técnicas de Data Mining a um SHAE, PCMAT - Mathematics Collaborative Educational System. A aplicação destas técnicas deram origem a modelos de dados que transformaram os dados em informações úteis e compreensíveis, essenciais para a geração de novos perfis de alunos, padrões de comportamento de alunos, regras de adaptação e pedagógicas. Neste trabalho foram criados alguns modelos de dados recorrendo à técnica de Data Mining de classificação, abordando diferentes algoritmos. Os resultados obtidos permitirão definir novas regras de adaptação e padrões de comportamento dos alunos, poderá melhorar o processo de aprendizagem disponível num SHAE.