Design and Evaluation of the Hamal Parallel Computer

Parallel shared-memory machines with hundreds or thousands of processor-memory nodes have been built; in the future we will see machines with millions or even billions of nodes. Associated with such large systems is a new set of design challenges. Many problems must be addressed by an architecture in order for it to be successful; of these, we focus on three in particular. First, a scalable memory system is required. Second, the network messaging protocol must be fault-tolerant. Third, the overheads of thread creation, thread management and synchronization must be extremely low. This thesis presents the complete system design for Hamal, a shared-memory architecture which addresses these concerns and is directly scalable to one million nodes. Virtual memory and distributed objects are implemented in a manner that requires neither inter-node synchronization nor the storage of globally coherent translations at each node. We develop a lightweight fault-tolerant messaging protocol that guarantees message delivery and idempotence across a discarding network. A number of hardware mechanisms provide efficient support for massive multithreading and fine-grained synchronization. Experiments are conducted in simulation, using a trace-driven network simulator to investigate the messaging protocol and a cycle-accurate simulator to evaluate the Hamal architecture. We determine implementation parameters for the messaging protocol which optimize performance. A discarding network is easier to design and can be clocked at a higher rate, and we find that with this protocol its performance can approach that of a non-discarding network. Our simulations of Hamal demonstrate the effectiveness of its thread management and synchronization primitives. In particular, we find register-based synchronization to be an extremely efficient mechanism which can be used to implement a software barrier with a latency of only 523 cycles on a 512 node machine.

Exogenous-Loss Awareness in Queue Management: Toward Global Fairness

For a given TCP flow, exogenous losses are those occurring on links other than the flow's bottleneck link. Exogenous losses are typically viewed as introducing undesirable "noise" into TCP's feedback control loop, leading to inefficient network utilization and potentially severe global unfairness. This has prompted much research on mechanisms for hiding such losses from end-points. In this paper, we show through analysis and simulations that low levels of exogenous losses are surprisingly beneficial in that they improve stability and convergence, without sacrificing efficiency. Based on this, we argue that exogenous loss awareness should be taken into account in any AQM design that aims to achieve global fairness. To that end, we propose an exogenous-loss aware Queue Management (XQM) that actively accounts for and leverages exogenous losses. We use an equation based approach to derive the quiescent loss rate for a connection based on the connection's profile and its global fair share. In contrast to other queue management techniques, XQM ensures that a connection sees its quiescent loss rate, not only by complementing already existing exogenous losses, but also by actively hiding exogenous losses, if necessary, to achieve global fairness. We establish the advantages of exogenous-loss awareness using extensive simulations in which, we contrast the performance of XQM to that of a host of traditional exogenous-loss unaware AQM techniques.

An energy-efficient data delivery scheme for delay-sensitive traffic in wireless sensor networks

We propose a novel data-delivery method for delay-sensitive traffic that significantly reduces the energy consumption in wireless sensor networks without reducing the number of packets that meet end-to-end real-time deadlines. The proposed method, referred to as SensiQoS, leverages the spatial and temporal correlation between the data generated by events in a sensor network and realizes energy savings through application-specific in-network aggregation of the data. SensiQoS maximizes energy savings by adaptively waiting for packets from upstream nodes to perform in-network processing without missing the real-time deadline for the data packets. SensiQoS is a distributed packet scheduling scheme, where nodes make localized decisions on when to schedule a packet for transmission to meet its end-to-end real-time deadline and to which neighbor they should forward the packet to save energy. We also present a localized algorithm for nodes to adapt to network traffic to maximize energy savings in the network. Simulation results show that SensiQoS improves the energy savings in sensor networks where events are sensed by multiple nodes, and spatial and/or temporal correlation exists among the data packets. Energy savings due to SensiQoS increase with increase in the density of the sensor nodes and the size of the sensed events. © 2010 Harshavardhan Sabbineni and Krishnendu Chakrabarty.

Residential electrical vehicle charging strategies: the good, the bad and the ugly

In recent years, a wide variety of centralised and decentralised algorithms have been proposed for residential charging of electric vehicles (EVs). In this paper, we present a mathematical framework which casts the EV charging scenarios addressed by these algorithms as optimisation problems having either temporal or instantaneous optimisation objectives with respect to the different actors in the power system. Using this framework and a realistic distribution network simulation testbed, we provide a comparative evaluation of a range of different residential EV charging strategies, highlighting in each case positive and negative characteristics.

Wireless networks congestion control

Congestion control in wireless networks is an important and open issue. Previous research has proven the poor performance of the Transport Control Protocol (TCP) in such networks. The factors that contribute to the poor performance of TCP in wireless environments concern its unsuitability to identify/detect and react properly to network events, its TCP window based ow control algorithm that is not suitable for the wireless channel, and the congestion collapse due to mobility. New rate based mechanisms have been proposed to mitigate TCP performance in wired and wireless networks. However, these mechanisms also present poor performance, as they lack of suitable bandwidth estimation techniques for multi-hop wireless networks. It is thus important to improve congestion control performance in wireless networks, incorporating components that are suitable for wireless environments. A congestion control scheme which provides an e - cient and fair sharing of the underlying network capacity and available bandwidth among multiple competing applications is crucial to the definition of new e cient and fair congestion control schemes on wireless multi-hop networks. The Thesis is divided in three parts. First, we present a performance evaluation study of several congestion control protocols against TCP, in wireless mesh and ad-hoc networks. The obtained results show that rate based congestion control protocols need an eficient and accurate underlying available bandwidth estimation technique. The second part of the Thesis presents a new link capacity and available bandwidth estimation mechanism denoted as rt-Winf (real time wireless inference). The estimation is performed in real-time and without the need to intrusively inject packets in the network. Simulation results show that rt-Winf obtains the available bandwidth and capacity estimation with accuracy and without introducing overhead trafic in the network. The third part of the Thesis proposes the development of new congestion control mechanisms to address the congestion control problems of wireless networks. These congestion control mechanisms use cross layer information, obtained by rt-Winf, to accurately and eficiently estimate the available bandwidth and the path capacity over a wireless network path. Evaluation of these new proposed mechanisms, through ns-2 simulations, shows that the cooperation between rt-Winf and the congestion control algorithms is able to significantly increase congestion control eficiency and network performance.

Energy and reserve provision dispatch considering distributed generation and demand response

In competitive electricity markets with deep concerns at the efficiency level, demand response programs gain considerable significance. In the same way, distributed generation has gained increasing importance in the operation and planning of power systems. Grid operators and utilities are taking new initiatives, recognizing the value of demand response and of distributed generation for grid reliability and for the enhancement of organized spot market´s efficiency. Grid operators and utilities become able to act in both energy and reserve components of electricity markets. This paper proposes a methodology for a joint dispatch of demand response and distributed generation to provide energy and reserve by a virtual power player that operates a distribution network. The proposed method has been computationally implemented and its application is illustrated in this paper using a 32 bus distribution network with 32 medium voltage consumers.

Intelligent energy resource management considering vehicle-to-grid: A simulated annealing approach

This paper proposes a simulated annealing (SA) approach to address energy resources management from the point of view of a virtual power player (VPP) operating in a smart grid. Distributed generation, demand response, and gridable vehicles are intelligently managed on a multiperiod basis according to V2G user´s profiles and requirements. Apart from using the aggregated resources, the VPP can also purchase additional energy from a set of external suppliers. The paper includes a case study for a 33 bus distribution network with 66 generators, 32 loads, and 1000 gridable vehicles. The results of the SA approach are compared with a methodology based on mixed-integer nonlinear programming. A variation of this method, using ac load flow, is also used and the results are compared with the SA solution using network simulation. The proposed SA approach proved to be able to obtain good solutions in low execution times, providing VPPs with suitable decision support for the management of a large number of distributed resources.

Demand response in electrical energy supply: an optimal real time pricing approach

In competitive electricity markets with deep concerns for the efficiency level, demand response programs gain considerable significance. As demand response levels have decreased after the introduction of competition in the power industry, new approaches are required to take full advantage of demand response opportunities. This paper presents DemSi, a demand response simulator that allows studying demand response actions and schemes in distribution networks. It undertakes the technical validation of the solution using realistic network simulation based on PSCAD. The use of DemSi by a retailer in a situation of energy shortage, is presented. Load reduction is obtained using a consumer based price elasticity approach supported by real time pricing. Non-linear programming is used to maximize the retailer’s profit, determining the optimal solution for each envisaged load reduction. The solution determines the price variations considering two different approaches, price variations determined for each individual consumer or for each consumer type, allowing to prove that the approach used does not significantly influence the retailer’s profit. The paper presents a case study in a 33 bus distribution network with 5 distinct consumer types. The obtained results and conclusions show the adequacy of the used methodology and its importance for supporting retailers’ decision making.

Demsi - a demand response simulator in the context of intensive use of distributed generation

Demand response can play a very relevant role in future power systems in which distributed generation can help to assure service continuity in some fault situations. This paper deals with the demand response concept and discusses its use in the context of competitive electricity markets and intensive use of distributed generation. The paper presents DemSi, a demand response simulator that allows studying demand response actions and schemes using a realistic network simulation based on PSCAD. Demand response opportunities are used in an optimized way considering flexible contracts between consumers and suppliers. A case study evidences the advantages of using flexible contracts and optimizing the available generation when there is a lack of supply.

Sobrevivência das redes de transporte da nova geração

Trabalho Final de Mestrado para obtenção do grau de Mestre em Engenharia de Electrónica e Telecomunicações

Dispatch of distributed Energy Resources to Provide Energy and Reserve in Smart Grids using a Particle Swarm Optimization Approach

The smart grid concept is a key issue in the future power systems, namely at the distribution level, with deep concerns in the operation and planning of these systems. Several advantages and benefits for both technical and economic operation of the power system and of the electricity markets are recognized. The increasing integration of demand response and distributed generation resources, all of them mostly with small scale distributed characteristics, leads to the need of aggregating entities such as Virtual Power Players. The operation business models become more complex in the context of smart grid operation. Computational intelligence methods can be used to give a suitable solution for the resources scheduling problem considering the time constraints. This paper proposes a methodology for a joint dispatch of demand response and distributed generation to provide energy and reserve by a virtual power player that operates a distribution network. The optimal schedule minimizes the operation costs and it is obtained using a particle swarm optimization approach, which is compared with a deterministic approach used as reference methodology. The proposed method is applied to a 33-bus distribution network with 32 medium voltage consumers and 66 distributed generation units.

Mobile Protocol Stack Simulator

Nykypäivän maailma tukeutuu verkkoihin. Tietokoneverkot ja langattomat puhelimet ovat jo varsin tavallisia suurelle joukolle ihmisiä. Uusi verkkotyyppi on ilmestynyt edelleen helpottamaan ihmisten verkottunutta elämää. Ad hoc –verkot mahdollistavat joustavan verkonmuodostuksen langattomien päätelaitteiden välille ilman olemassa olevaa infrastruktuuria. Diplomityö esittelee uuden simulaatiotyökalun langattomien ad hoc –verkkojen simulointiin protokollatasolla. Se esittelee myös kyseisten verkkojen taustalla olevat periaatteet ja teoriat. Lähemmin tutkitaan OSI-mallin linkkikerroksen kaistanjakoprotokollia ad hoc –verkoissa sekä vastaavan toteutusta simulaattorissa. Lisäksi esitellään joukko simulaatioajoja esimerkiksi simulaattorin toiminnasta ja mahdollisista käyttökohteista.

Thread Scheduling Mechanisms for Multiple-Context Parallel Processors

Scheduling tasks to efficiently use the available processor resources is crucial to minimizing the runtime of applications on shared-memory parallel processors. One factor that contributes to poor processor utilization is the idle time caused by long latency operations, such as remote memory references or processor synchronization operations. One way of tolerating this latency is to use a processor with multiple hardware contexts that can rapidly switch to executing another thread of computation whenever a long latency operation occurs, thus increasing processor utilization by overlapping computation with communication. Although multiple contexts are effective for tolerating latency, this effectiveness can be limited by memory and network bandwidth, by cache interference effects among the multiple contexts, and by critical tasks sharing processor resources with less critical tasks. This thesis presents techniques that increase the effectiveness of multiple contexts by intelligently scheduling threads to make more efficient use of processor pipeline, bandwidth, and cache resources. This thesis proposes thread prioritization as a fundamental mechanism for directing the thread schedule on a multiple-context processor. A priority is assigned to each thread either statically or dynamically and is used by the thread scheduler to decide which threads to load in the contexts, and to decide which context to switch to on a context switch. We develop a multiple-context model that integrates both cache and network effects, and shows how thread prioritization can both maintain high processor utilization, and limit increases in critical path runtime caused by multithreading. The model also shows that in order to be effective in bandwidth limited applications, thread prioritization must be extended to prioritize memory requests. We show how simple hardware can prioritize the running of threads in the multiple contexts, and the issuing of requests to both the local memory and the network. Simulation experiments show how thread prioritization is used in a variety of applications. Thread prioritization can improve the performance of synchronization primitives by minimizing the number of processor cycles wasted in spinning and devoting more cycles to critical threads. Thread prioritization can be used in combination with other techniques to improve cache performance and minimize cache interference between different working sets in the cache. For applications that are critical path limited, thread prioritization can improve performance by allowing processor resources to be devoted preferentially to critical threads. These experimental results show that thread prioritization is a mechanism that can be used to implement a wide range of scheduling policies.

On social networks and game theory, some research problems

Seminar given as part of social networking course, to give a brief overview of some applied examples game theory used in social network simulation

Técnicas de avaliação da confiabilidade em estruturas de concreto armado

Neste trabalho é dado ênfase à inclusão das incertezas na avaliação do comportamento estrutural, objetivando uma melhor representação das características do sistema e uma quantificação do significado destas incertezas no projeto. São feitas comparações entre as técnicas clássicas existentes de análise de confiabilidade, tais como FORM, Simulação Direta Monte Carlo (MC) e Simulação Monte Carlo com Amostragem por Importância Adaptativa (MCIS), e os métodos aproximados da Superfície de Resposta( RS) e de Redes Neurais Artificiais(ANN). Quando possível, as comparações são feitas salientando- se as vantagens e inconvenientes do uso de uma ou de outra técnica em problemas com complexidades crescentes. São analisadas desde formulações com funções de estado limite explícitas até formulações implícitas com variabilidade espacial de carregamento e propriedades dos materiais, incluindo campos estocásticos. É tratado, em especial, o problema da análise da confiabilidade de estruturas de concreto armado incluindo o efeito da variabilidade espacial de suas propriedades. Para tanto é proposto um modelo de elementos finitos para a representação do concreto armado que incorpora as principais características observadas neste material. Também foi desenvolvido um modelo para a geração de campos estocásticos multidimensionais não Gaussianos para as propriedades do material e que é independente da malha de elementos finitos, assim como implementadas técnicas para aceleração das avaliações estruturais presentes em qualquer das técnicas empregadas. Para o tratamento da confiabilidade através da técnica da Superfície de Resposta, o algoritmo desenvolvido por Rajashekhar et al(1993) foi implementado. Já para o tratamento através de Redes Neurais Artificias, foram desenvolvidos alguns códigos para a simulação de redes percéptron multicamada e redes com função de base radial e então implementados no algoritmo de avaliação de confiabilidade desenvolvido por Shao et al(1997). Em geral, observou-se que as técnicas de simulação tem desempenho bastante baixo em problemas mais complexos, sobressaindo-se a técnica de primeira ordem FORM e as técnicas aproximadas da Superfície de Resposta e de Redes Neurais Artificiais, embora com precisão prejudicada devido às aproximações presentes.