On cluster analysis of complex and heterogeneous data

3rd SMTDA Conference Proceedings, 11-14 June 2014, Lisbon Portugal.

Optimal virtual cluster-based multiprocessor scheduling

Scheduling of constrained deadline sporadic task systems on multiprocessor platforms is an area which has received much attention in the recent past. It is widely believed that finding an optimal scheduler is hard, and therefore most studies have focused on developing algorithms with good processor utilization bounds. These algorithms can be broadly classified into two categories: partitioned scheduling in which tasks are statically assigned to individual processors, and global scheduling in which each task is allowed to execute on any processor in the platform. In this paper we consider a third, more general, approach called cluster-based scheduling. In this approach each task is statically assigned to a processor cluster, tasks in each cluster are globally scheduled among themselves, and clusters in turn are scheduled on the multiprocessor platform. We develop techniques to support such cluster-based scheduling algorithms, and also consider properties that minimize total processor utilization of individual clusters. In the last part of this paper, we develop new virtual cluster-based scheduling algorithms. For implicit deadline sporadic task systems, we develop an optimal scheduling algorithm that is neither Pfair nor ERfair. We also show that the processor utilization bound of us-edf{m/(2m−1)} can be improved by using virtual clustering. Since neither partitioned nor global strategies dominate over the other, cluster-based scheduling is a natural direction for research towards achieving improved processor utilization bounds.

TDBS: a time division beacon scheduling mechanism for ZigBee cluster-tree wireless sensor networks

Synchronization is a challenging and important issue for time-sensitive Wireless Sensor Networks (WSN) since it requires a mutual spatiotemporal coordination between the nodes. In that concern, the IEEE 802.15.4/ZigBee protocols embody promising technologies for WSNs, but are still ambiguous on how to efficiently build synchronized multiple-cluster networks, specifically for the case of cluster-tree topologies. In fact, the current IEEE 802.15.4/ZigBee specifications restrict the synchronization to beacon-enabled (by the generation of periodic beacon frames) star networks, while they support multi-hop networking in mesh topologies, but with no synchronization. Even though both specifications mention the possible use of cluster-tree topologies, which combine multi-hop and synchronization features, the description on how to effectively construct such a network topology is missing. This paper tackles this issue by unveiling the ambiguities regarding the use of the cluster-tree topology and proposing a synchronization mechanism based on Time Division Beacon Scheduling (TDBS) to build cluster-tree WSNs. In addition, we propose a methodology for efficiently managing duty-cycles in every cluster, ensuring the fairest use of bandwidth resources. The feasibility of the TDBS mechanism is clearly demonstrated through an experimental test-bed based on our open-source implementation of the IEEE 802.15.4/ZigBee protocols.

A time division beacon scheduling mechanism for IEEE 802.15.4/Zigbee cluster-tree wireless sensor networks

While the IEEE 802.15.4/Zigbee protocol stack is being considered as a promising technology for low-cost low-power Wireless Sensor Networks (WSNs), several issues in the standard specifications are still open. One of those ambiguous issues is how to build a synchronized multi-hop cluster-tree network, which is quite suitable for ensuring QoS support in WSNs. In fact, the current IEEE 802.15.4/Zigbee specifications restrict the synchronization in the beacon-enabled mode (by the generation of periodic beacon frames) to star-based networks, while it supports multi-hop networking using the peer-to-peer mesh topology, but with no synchronization. Even though both specifications mention the possible use of cluster-tree topologies, which combine multihop and synchronization features, the description on how to effectively construct such a network topology is missing. This paper tackles this problem, unveils the ambiguities regarding the use of the cluster-tree topology and proposes a synchronization mechanism based on Time Division Beacon Scheduling to construct cluster-tree WSNs. We also propose a methodology for an efficient duty cycle management in each router (cluster-head) of a cluster-tree WSN that ensures the fairest use of bandwidth resources. The feasibility of the proposal is clearly demonstrated through an experimental test bed based on our own implementation of the IEEE 802.15.4/Zigbee protocol.

Collision-free beacon scheduling mechanisms for IEEE 802.15.4/Zigbee cluster-tree wireless sensor networks

The recently standardized IEEE 802.15.4/Zigbee protocol stack offers great potentials for ubiquitous and pervasive computing, namely for Wireless Sensor Networks (WSNs). However, there are still some open and ambiguous issues that turn its practical use a challenging task. One of those issues is how to build a synchronized multi-hop cluster-tree network, which is quite suitable for QoS support in WSNs. In fact, the current IEEE 802.15.4/Zigbee specifications restrict the synchronization in the beacon-enabled mode (by the generation of periodic beacon frames) to star-based networks, while it supports multi-hop networking using the peer-to-peer mesh topology, but with no synchronization. Even though both specifications mention the possible use of cluster-tree topologies, which combine multi-hop and synchronization features, the description on how to effectively construct such a network topology is missing. This paper tackles this problem, unveils the ambiguities regarding the use of the cluster-tree topology and proposes two collision-free beacon frame scheduling schemes. We strongly believe that the results provided in this paper trigger a significant step towards the practical and efficient use of IEEE 802.15.4/Zigbee cluster-tree networks.

Beacon scheduling in cluster-tree IEEE 802.15.4/ZigBee wireless sensor networks

The recently standardized IEEE 802.15.4/Zigbee protocol stack offers great potentials for ubiquitous and pervasive computing, namely for Wireless Sensor Networks (WSNs). However, there are still some open and ambiguous issues that turn its practical use a challenging task. One of those issues is how to build a synchronized multi-hop cluster-tree network, which is quite suitable for QoS support in WSNs. In fact, the current IEEE 802.15.4/Zigbee specifications restrict the synchronization in the beacon-enabled mode (by the generation of periodic beacon frames) to star-based networks, while it supports multi-hop networking using the peer-to-peer mesh topology, but with no synchronization. Even though both specifications mention the possible use of cluster-tree topologies, which combine multi-hop and synchronization features, the description on how to effectively construct such a network topology is missing. This report tackles this problem, unveils the ambiguities regarding the use of the cluster-tree topology and proposes two collisionfree beacon frame scheduling schemes.

On the choice of posittioning and cluster size for fixed relay stations in a LTE network

The long term evolution (LTE) is one of the latest standards in the mobile communications market. To achieve its performance, LTE networks use several techniques, such as multi-carrier technique, multiple-input-multiple-output and cooperative communications. Inside cooperative communications, this paper focuses on the fixed relaying technique, presenting a way for determining the best position to deploy the relay station (RS), from a set of empirical good solutions, and also to quantify the associated performance gain using different cluster size configurations. The best RS position was obtained through realistic simulations, which set it as the middle of the cell's circumference arc. Additionally, it also confirmed that network's performance is improved when the number of RSs is increased. It was possible to conclude that, for each deployed RS, the percentage of area served by an RS increases about 10 %. Furthermore, the mean data rate in the cell has been increased by approximately 60 % through the use of RSs. Finally, a given scenario with a larger number of RSs, can experience the same performance as an equivalent scenario without RSs, but with higher reuse distance. This conduces to a compromise solution between RS installation and cluster size, in order to maximize capacity, as well as performance.

Protocolo Mac para acesso multi-modo em redes de sensores sem fios móveis

Dissertação apresentada na Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa para a obtenção do grau de Mestre em Engenharia Electrotécnica e de Computadores

Optical processor based on a-SiC technology for spectral data error control

The SiC optical processor for error detection and correction is realized by using double pin/pin a-SiC:H photodetector with front and back biased optical gating elements. Data shows that the background act as selector that pick one or more states by splitting portions of the input multi optical signals across the front and back photodiodes. Boolean operations such as exclusive OR (EXOR) and three bit addition are demonstrated optically with a combination of such switching devices, showing that when one or all of the inputs are present the output will be amplified, the system will behave as an XOR gate representing the SUM. When two or three inputs are on, the system acts as AND gate indicating the present of the CARRY bit. Additional parity logic operations are performed by use of the four incoming pulsed communication channels that are transmitted and checked for errors together. As a simple example of this approach, we describe an all optical processor for error detection and correction and then, provide an experimental demonstration of this fault tolerant reversible system, in emerging nanotechnology.

Multi-GPU support on the marrow algorithmic skeleton framework

Dissertação para obtenção do Grau de Mestre em Engenharia Informática

Towards an algorithmic skeleton framework for programming the Intel R Xeon PhiTM processor

The Intel R Xeon PhiTM is the first processor based on Intel’s MIC (Many Integrated Cores) architecture. It is a co-processor specially tailored for data-parallel computations, whose basic architectural design is similar to the ones of GPUs (Graphics Processing Units), leveraging the use of many integrated low computational cores to perform parallel computations. The main novelty of the MIC architecture, relatively to GPUs, is its compatibility with the Intel x86 architecture. This enables the use of many of the tools commonly available for the parallel programming of x86-based architectures, which may lead to a smaller learning curve. However, programming the Xeon Phi still entails aspects intrinsic to accelerator-based computing, in general, and to the MIC architecture, in particular. In this thesis we advocate the use of algorithmic skeletons for programming the Xeon Phi. Algorithmic skeletons abstract the complexity inherent to parallel programming, hiding details such as resource management, parallel decomposition, inter-execution flow communication, thus removing these concerns from the programmer’s mind. In this context, the goal of the thesis is to lay the foundations for the development of a simple but powerful and efficient skeleton framework for the programming of the Xeon Phi processor. For this purpose we build upon Marrow, an existing framework for the orchestration of OpenCLTM computations in multi-GPU and CPU environments. We extend Marrow to execute both OpenCL and C++ parallel computations on the Xeon Phi. We evaluate the newly developed framework, several well-known benchmarks, like Saxpy and N-Body, will be used to compare, not only its performance to the existing framework when executing on the co-processor, but also to assess the performance on the Xeon Phi versus a multi-GPU environment.

Long-term effect of a school-based physical activity program (KISS) on fitness and adiposity in children: a cluster-randomized controlled trial.

BACKGROUND: School-based intervention studies promoting a healthy lifestyle have shown favorable immediate health effects. However, there is a striking paucity on long-term follow-ups. The aim of this study was therefore to assess the 3 yr-follow-up of a cluster-randomized controlled school-based physical activity program over nine month with beneficial immediate effects on body fat, aerobic fitness and physical activity. METHODS AND FINDINGS: Initially, 28 classes from 15 elementary schools in Switzerland were grouped into an intervention (16 classes from 9 schools, n = 297 children) and a control arm (12 classes from 6 schools, n = 205 children) after stratification for grade (1st and 5th graders). Three years after the end of the multi-component physical activity program of nine months including daily physical education (i.e. two additional lessons per week on top of three regular lessons), short physical activity breaks during academic lessons, and daily physical activity homework, 289 (58%) participated in the follow-up. Primary outcome measures included body fat (sum of four skinfolds), aerobic fitness (shuttle run test), physical activity (accelerometry), and quality of life (questionnaires). After adjustment for grade, gender, baseline value and clustering within classes, children in the intervention arm compared with controls had a significantly higher average level of aerobic fitness at follow-up (0.373 z-score units [95%-CI: 0.157 to 0.59, p = 0.001] corresponding to a shift from the 50th to the 65th percentile between baseline and follow-up), while the immediate beneficial effects on the other primary outcomes were not sustained. CONCLUSIONS: Apart from aerobic fitness, beneficial effects seen after one year were not maintained when the intervention was stopped. A continuous intervention seems necessary to maintain overall beneficial health effects as reached at the end of the intervention. TRIAL REGISTRATION: ControlledTrials.com ISRCTN15360785.

Análisis de prestaciones de procesadores multi-core y multi-thread

Los procesadores multi-core y el multi-threading por hardware permiten aumentar el rendimiento de las aplicaciones. Por un lado, los procesadores multi-core combinan 2 o más procesadores en un mismo chip. Por otro lado, el multi-threading por hardware es una técnica que incrementa la utilización de los recursos del procesador. Este trabajo presenta un análisis de rendimiento de los resultados obtenidos en dos aplicaciones, multiplicación de matrices densas y transformada rápida de Fourier. Ambas aplicaciones se han ejecutado en arquitecturas multi-core que explotan el paralelismo a nivel de thread pero con un modelo de multi-threading diferente. Los resultados obtenidos muestran la importancia de entender y saber analizar el efecto del multi-core y multi-threading en el rendimiento.

Gestión de recursos en nodos multi-core de memoria compartida

La gestión de recursos en los procesadores multi-core ha ganado importancia con la evolución de las aplicaciones y arquitecturas. Pero esta gestión es muy compleja. Por ejemplo, una misma aplicación paralela ejecutada múltiples veces con los mismos datos de entrada, en un único nodo multi-core, puede tener tiempos de ejecución muy variables. Hay múltiples factores hardware y software que afectan al rendimiento. La forma en que los recursos hardware (cómputo y memoria) se asignan a los procesos o threads, posiblemente de varias aplicaciones que compiten entre sí, es fundamental para determinar este rendimiento. La diferencia entre hacer la asignación de recursos sin conocer la verdadera necesidad de la aplicación, frente a asignación con una meta específica es cada vez mayor. La mejor manera de realizar esta asignación és automáticamente, con una mínima intervención del programador. Es importante destacar, que la forma en que la aplicación se ejecuta en una arquitectura no necesariamente es la más adecuada, y esta situación puede mejorarse a través de la gestión adecuada de los recursos disponibles. Una apropiada gestión de recursos puede ofrecer ventajas tanto al desarrollador de las aplicaciones, como al entorno informático donde ésta se ejecuta, permitiendo un mayor número de aplicaciones en ejecución con la misma cantidad de recursos. Así mismo, esta gestión de recursos no requeriría introducir cambios a la aplicación, o a su estrategia operativa. A fin de proponer políticas para la gestión de los recursos, se analizó el comportamiento de aplicaciones intensivas de cómputo e intensivas de memoria. Este análisis se llevó a cabo a través del estudio de los parámetros de ubicación entre los cores, la necesidad de usar la memoria compartida, el tamaño de la carga de entrada, la distribución de los datos dentro del procesador y la granularidad de trabajo. Nuestro objetivo es identificar cómo estos parámetros influyen en la eficiencia de la ejecución, identificar cuellos de botella y proponer posibles mejoras. Otra propuesta es adaptar las estrategias ya utilizadas por el Scheduler con el fin de obtener mejores resultados.

Effect of school based physical activity programme (KISS) on fitness and adiposity in primary schoolchildren: cluster randomised controlled trial.

To assess the effectiveness of a school based physical activity programme during one school year on physical and psychological health in young schoolchildren. Cluster randomised controlled trial. 28 classes from 15 elementary schools in Switzerland randomly selected and assigned in a 4:3 ratio to an intervention (n=16) or control arm (n=12) after stratification for grade (first and fifth grade), from August 2005 to June 2006. 540 children, of whom 502 consented and presented at baseline. Children in the intervention arm (n=297) received a multi-component physical activity programme that included structuring the three existing physical education lessons each week and adding two additional lessons a week, daily short activity breaks, and physical activity homework. Children (n=205) and parents in the control group were not informed of an intervention group. For most outcome measures, the assessors were blinded. Primary outcome measures included body fat (sum of four skinfolds), aerobic fitness (shuttle run test), physical activity (accelerometry), and quality of life (questionnaires). Secondary outcome measures included body mass index and cardiovascular risk score (average z score of waist circumference, mean blood pressure, blood glucose, inverted high density lipoprotein cholesterol, and triglycerides). 498 children completed the baseline and follow-up assessments (mean age 6.9 (SD 0.3) years for first grade, 11.1 (0.5) years for fifth grade). After adjustment for grade, sex, baseline values, and clustering within classes, children in the intervention arm compared with controls showed more negative changes in the z score of the sum of four skinfolds (-0.12, 95 % confidence interval -0.21 to -0.03; P=0.009). Likewise, their z scores for aerobic fitness increased more favourably (0.17, 0.01 to 0.32; P=0.04), as did those for moderate-vigorous physical activity in school (1.19, 0.78 to 1.60; P<0.001), all day moderate-vigorous physical activity (0.44, 0.05 to 0.82; P=0.03), and total physical activity in school (0.92, 0.35 to 1.50; P=0.003). Z scores for overall daily physical activity (0.21, -0.21 to 0.63) and physical quality of life (0.42, -1.23 to 2.06) as well as psychological quality of life (0.59, -0.85 to 2.03) did not change significantly. A school based multi-component physical activity intervention including compulsory elements improved physical activity and fitness and reduced adiposity in children. Trial registration Current Controlled Trials ISRCTN15360785.