Aravind Chandrashekar, for the Master of Science degree in Electrical and Computer, presented on 02/09/2011, at Southern Illinois University Carbondale. TITLE: Reduction of Cache Related Preemption Delay using DVS in Real Time Systems MAJOR PROFESSOR: Dr. Harini Ramaprasad Embedded/real-time systems are ubiquitous in today's world. Providing temporal guarantees is paramount in such systems. In several multi-tasking real-time systems, tasks are assigned varying priorities and scheduled in accordance with a preemptive scheduling policy. When a task is preempted, a significant number of memory blocks belonging to the particular task are displaced from the cache memory between the time that the task is preempted and the time that the task resumes execution. Upon resumption, a corresponding amount of time is spent in reloading the cache with previously replaced memory blocks, thereby incurring what is known as cache-related preemption delay (CRPD). CRPD of a task due to a given preemption depends on the position in the program where the preempted task is executing at the time of preemption. As such, CRPD at different preemption points may be significantly different. In this thesis, we exploit this difference in CRPD and use dynamic voltage/frequency scaling (DVFS) to control the execution speed of a task such that it gets preempted in regions where the CRPD is low, as far as is possible without jeopardizing system schedulability. Simulation results demonstrate that our algorithm reduces number of cache reloads due to preemption to a reasonable extent, thereby reducing the repeated usage of off-chip memory bandwidth.

Dissertação (mestrado) - Universidade Federal de Santa Catarina, Centro Tecnológico. Programa de Pós-Graduação em Engenharia de Automação e Sistemas / Made available in DSpace on 2012-10-26T12:25:42Z (GMT). No. of bitstreams: 1 301047.pdf: 1285723 bytes, checksum: fcb30ba5e90539742c71505b32d65921 (MD5) / Sistemas de tempo real são sistemas onde o correto funcionamento não depende somente da resposta lógica correta, mas também do tempo no qual ela foi dada. Igualmente do ponto de vista lógico, a viabilidade temporal da aplicação deve ser determinada através de técnicas, como por exemplo análise do tempo de reposta. Este tipo de aplicação está cada vez mais presente atualmente e a demanda de processamento é tamanha que necessita-se de processadores com múltiplos núcleos complexos. É perceptível que o desenvolvimento dos multiprocessadores está muito mais avançado em relação às técnicas de análise de tais sistemas e, portanto, é evidente a necessidade de pesquisa com objetivo de promover maior confiabilidade e redução de superdimensionamentos. O objetivo deste trabalho é promover uma solução de escalonamento que considere a escalonabilidade em conjunto com a analisabilidade do código da aplicação. Atualmente, a pesquisa de sistemas de tempo real trata o problema do escalonamento isolado do problema de obtenção do parâmetro do tempo de computação da tarefas (WCET --Worst Case Execution Time). Dependendo da arquitetura do processador, as premissas adotadas no cálculo do WCET são incompatíveis com as premissas de escalonamento, o que gera uma contradição fundamental entre o cálculo do WCET e os algoritmos de escalonamento. A incompatibilidade das premissas pode ser ilustrada pela preempção em arquiteturas com memória cache, onde o cálculo de WCET assume execução contínua da tarefa, o que não é verdade em grande parte dos algoritmos de escalonamento. Este trabalho propõe o uso de uma abordagem heterogênea em multiprocessadores onde parte dos núcleos operam em regime preemptivo e parte em regime não preemptivo para tentar lidar com as diferentes considerações sobre preempção. As análises realizadas mostram que existe vantagem em usar a abordagem heterogênea. / Real-time systems are systems where the correct functioning depends not only on the logically correct response, but also the time when it was given. As the the logic functionality, the application response time could be analyzed to determine the viability of a real-time system. This type of application is increasingly present today and the processing demand is such that complex multi-core processors are needed. It is noticeable that the development of multiprocessor is a long way ahead compared with the techniques of analysis of such systems and is therefore necessary researches to promote more reliability and to reduce over-specified systems. The objective of this work is to promote a solution that considers scheduling in conjunction with the analyzability of the application code. Currently, the real-time research considers the scheduling problem isolated from the WCET (Worst Case Execution Time) problem. Depending on the processor architecture, the values obtained by computing WCET are incompatible with the scheduling model which creates a fundamental contradiction between the assumptions of calculation of WCET and scheduling algorithms. This work proposes the use of a heterogeneous approach where part of the multiprocessor cores operate under preemptive and part on a non-preemptive scheduling. The analysis shows that there are advantages using the heterogeneous approach.

Dissertação (mestrado) - Universidade Federal de Santa Catarina, Centro Tecnológico. Programa de Pós-Graduação em Ciência da Computação. / Made available in DSpace on 2012-10-21T06:00:21Z (GMT). No. of bitstreams: 0 / O presente trabalho tem como objetivo a avaliação do desempenho de uma Hierarquia de Proxies, dedicada a Cache de Web, localizada em redes de alta velocidade. A análise foi realizada dando enfoque ao tempo de resposta para o usuário final propriamente dito. Verificou-se, utilizando máquinas clientes, que em uma rede de alta velocidade onde se localiza uma Hierarquia de Proxies, os tempos de respostas apresentaram variações favoráveis ao uso dos Proxies. Entretanto verificou-se que a implementação deste serviço, neste tipo de rede, não apresenta ganhos muitos significativos no que diz respeito aos tempos de resposta para o usuário final, ficando em torno de 1 segundo. Constatou-se também que o tempo gasto na resolução de nomes das URLs pode degradar a performance dos acessos às páginas Web, seja utilizando hierarquia de Proxies ou nos simples acessos cliente-servidor. O problema foi verificado nas URLs que possuem registros DNS com baixo tempo vida (TTL).

Dissertação (mestrado)—Universidade de Brasília, Instituto de Ciências Exatas, Departamento de Ciência da Computação, 2016. / Submitted by Fernanda Percia França (fernandafranca@bce.unb.br) on 2017-03-10T17:17:48Z No. of bitstreams: 1 2016_CarlosBotelhodePaulaFilho.pdf: 3959000 bytes, checksum: cc882cfd3824c33e2ae1cd9d7dd0d0fa (MD5) / Approved for entry into archive by Raquel Viana(raquelviana@bce.unb.br) on 2017-04-04T21:04:09Z (GMT) No. of bitstreams: 1 2016_CarlosBotelhodePaulaFilho.pdf: 3959000 bytes, checksum: cc882cfd3824c33e2ae1cd9d7dd0d0fa (MD5) / Made available in DSpace on 2017-04-04T21:04:09Z (GMT). No. of bitstreams: 1 2016_CarlosBotelhodePaulaFilho.pdf: 3959000 bytes, checksum: cc882cfd3824c33e2ae1cd9d7dd0d0fa (MD5) / Cache na internet e um mecanismo chave para o aumento de performance e estabilidade de sites em momentos de grande número de acessos. Entretanto a maioria das estratégias de cache não levam em conta o poder do cache localizado entre nós clien- tes. Novas tecnologias desenvolvidas para a nova versão da especificação do HTML, o HTML5 possibilitaram que nós se conectem e interajam entre si, possibilitando a troca de dados sem a instalação de plugins ou addons de navegadores. Este trabalho descreve uma estratégia de cache que cria uma rede P2P de clientes agindo como servidores de proxy cache assim que os mesmos fizerem a primeira requisição ao servidor, distribuindo os dados potencialmente entre todos os visitantes do site. Aumentando a disponibilidade do site em picos de acesso, e minimizando drasticamente os custos de infraestrutura, e como os nós irão prover conteúdo dentro das redes dos provedores de internet, a estratégia pode potencialmente aumentar a velocidade de abertura de página ou de conteúdos de mídia. Assim como uma diminuição no custo de transferência de dados nos provedores de internet. / Web caching has been a key player in increasing performance and website stability in times of heavy usage. However most caching strategies do not take into account the power of localized caching between client nodes. New technologies developed for the new version of HTML specification, HTML5 have enabled nodes to connect and interact with each other, enabling them to share data without the addition of browser plugins or add-ons. This work describes a caching strategy that creates a P2P network of clients that act as proxy cache servers as soon as they issue a request to the server, distributing the caching data potentially across all web site visitors. Increasing the web site availability in high peaks, minimizing drastically the costs of infrastructure and, since the nodes will provide content to localized clients within their ISP networks, the strategy will potentially result in a faster overall speed in page loads, or in media content loading. As well as decreasing costs for ISPs due to the minimization of data exchange outside of the ISP network.

The goal of diploma thesis is disquisition on existing file format intended for data transmission over computer network. Choose one of the compared file format and create a file structure for transmission sports data. Assess the structure of files from existing sports data suppliers and create their own structure. Compare the possible ways of distribution of created data files and give recommendations for services implementers.

Com a grande quantidade de informações existentes nas aplicações computacionais hoje em dia, cada vez mais tornam-se necessários mecanismos que facilitem e aumentem o desempenho da recuperação dessas informações. Nesse contexto vem surgindo os bancos de dados chamados de NOSQL, que são bancos de dados tipicamente não relacionais que, em prol da disponibilidade e do desempenho em ambientes com enormes quantidades de dados, abrem mão de requisitos antes vistos como fundamentais. Neste trabalho iremos lidar com esse cenário ao implementar o módulo de consultas distribuídas do JBoss Infinispan, um sistema de cache distribuído que funciona também como um banco de dados NOSQL em memória. Além de apresentar a implementação desse módulo, iremos falar do surgimento do movimento NOSQL, de como se caracterizam esses bancos e de onde o Infinispan se insere nesse movimento. / With the big amount of data available to computer applications nowadays, there is an increasing need for mechanisms that facilitate the retrieval of such data and improve data access performance. In this context we see the emergence of so-called NOSQL databases, which are databases that are typically non-relational and that give up fulfilling some requirements previously seen as fundamental in order to achieve better availability and performance in big data environments. In this work we deal with the scenario above and implement a module that supports distributed queries in JBoss Infinispan, a distributed cache system that works also as an in-memory NOSQL database. Besides presenting the implementation of that module, we discuss the emergence of the NOSQL movement, the characterization of NOSQL databases, and where Infinispan fits in this context.

The importance of a smooth user experience in applications is increasing. To achieve more performance when interacting with resource intensive data it is important to implement an efficient caching method. The goal of this thesis is to investigate how to implement an efficient cache in an Android application. The use case is to download metadata and images of movies from a WebAPI provided by June AB. In order to investigate which caching method is the most efficient, a pre-study was done on some of the most common caching methods today. Based on the results of the pre-study, two different caching algorithms were tested and evaluated: First-In First-Out (FIFO) and Least Recently Used (LRU). These two algorithms were then implemented in an Android application. The resulting prototype has a responsive user interface capable of caching large amounts of data without noticeable performance loss compared to a non-cached version. The results from the prototype showed that LRU is the better strategy in our use case, however what we discovered was that the buffer size of the cache has the biggest impact on performance, not the cache eviction strategy. / Vikten av en snabb användarupplevelse ökar i nya applikationer. För att få ut mer prestanda när användare interagerar med resurstung data är det viktigt att implementera en effektiv cachingsmetod. Målet med arbetet är att undersöka hur man implementerar en effektiv cache i en Android-applikation. Användarfallet är att ladda ner metadata och bilder på filmer från ett WebAPI som tillhandahölls av June AB. För att undersöka vilken cachingsmetod som är effektivast gjordes en förstudie på några av de mest vanliga cachingsmetoderna idag. Baserat på förstudiens resultat valdes två cachingsalgoritmer för testning och utvärdering: First-In First-Out (FIFO) och Least Recently Used (LRU). Dessa två algoritmer implementerades i en Android-applikation Prototypen som gjordes har ett responsivt användargränsnitt som kan cacha stora mängder data utan märkbar prestandaförlust jämfört med en icke-cachad version. Prototypen visade att LRU är den bättre strategin för vårt användarfall, men upptäckte att bufferstorleken på cachen har den största påverkan av prestandan, inte cachestrategin.

Det är viktigare än någonsin att optimera svarstiden för databasförfrågningardå internettrafiken ökar och storleken på data växer. IT-företag har också blivitmer medvetna om vikten av att snabbt leverera innehåll till slutanvändaren pågrund av hur långsammare svarstider kan påverka kvalitetsuppfattningen påen produkt/ett system. Detta kan i sin tur leda till en negativ påverkan på ettföretags intäkter.I det här arbetet utvecklar och implementerar vi en ny dynamisk cachelösningpå applikationssidan av systemet och testar den mot väletablerade cachestrategier. Vi undersökte kända cache-strategier och relaterad forskning somtar hänsyn till den aktuella databasbelastningen så som historisk frekvens fören specifik databasförfrågan och tillämpade detta i vår algoritm. Vi utveckladefrån detta en dynamisk cachepolicy som använder en logaritmisk beräkningsom involverar den historiska frekvensen tillsammans med endatabasförfrågans svarstid och beräknade en vikt för en viss databasförfrågan.Vikten ger sedan prioritet i förhållande till andra databasförfrågningar som ärcachade. Vi kan här påvisa en prestandahöjning på 11-12% mot LRU, enprestandahöjning på 15% mot FIFO och en väsentlig prestandahöjning mot attanvända databasen direkt med både MySQL-cache aktiverad och inaktiverad. / With IP traffic and data sets continuously growing together with IT companiesbecoming more and more dependent on large data sets, it is more importantthan ever to optimize the load time of queries. IT companies have also becomemore aware of the importance of delivering content quickly to the end userbecause of how slower response times can affect quality perception which inturn can have a negative impact on revenue.In this paper, we develop and implement a new dynamic cache managementsystem with the cache on the application side of the system and test it againstwell-established caching policies. By looking at known caching strategies andresearch that takes the current database load into account with attributes suchas a queries frequency and incorporating this into our algorithm, we developeda dynamic caching policy that utilizes a logarithmic calculation involvinghistorical query frequency together with query response time to calculate aweight for a specific query. The weight gives priority in relation to other queriesresiding within the cache, which shows a performance increase towards existingcaching policies. The results show that we have a 11-12 % performance increasetowards LRU, a 15 % performance increase towards FIFO and a substantialperformance increase towards using the database directly with both MySQLcaching enabled and disabled.

Student understanding and use of tobacco was studied under t he following specific areas: specific knowledge as it relates to smoking and health; student opinions and attitudes toward smoking; smoking experience; and the influence of certain factors on smoking status. About 700 students were sampled with the use of a questionnaire. The number of students was divided about equally among four schools and between boys and girls at each grade level, grades seven through twelve. It was found t hat boys were better informed than girls about tobacco as it relates to health. Also, the senior high school students were better informed than the junior high school students. The majority (91 percent) of all students surveyed were of the opinion that the pleasure derived from smoking was not worth the price a person has to pay in terms of health and expense. Few students like having their parents smoke, and most are violently opposed to having them smoke. There were nine percent of the students who smoked regularly (at least once a week) with another eight percent smoking once in a while. This ranged from three percent in the seventh grade to fifteen percent in the twelfth grade. The factors having the most influence upon student smoking status were religion, friends, and health implications.

A comparative study of phytoseiid populations was made of two well- cultivated and regularly sprayed apple orchards with two unsprayed orchards in Cache Valley, northern Utah. Two phytoseiid species, Typhlodromus mcgregori Chant and T. occidentalis Nesbitt were observed on the apple leaves, under the bark, and occasionally in the litter and soil. Amblyseius cucumeris (Oudemans) occurred in the soil and litter and occasionally under bark. T. mcgregori was dominant in the unsprayed orchards, and T. occidentalis in the sprayed. Of the phytophagous mites, which served as food for the phytoseiids, the two-spotted mite, Tetranychus urticae Koch was dominant. Other phytophagous mites were the brown mite, Bryobia rubrioculus (Scheuten), the European red mite, Panonychus ulmi (Koch), and the McDaniel mite, Tetranychus mcdanieli McGregor. The study suggests T. mcgregori to be a non-specific, facultative predator of phytophagous mites. T. mcgregori was adversely affected by standard pesticide practices, but T. occiden talis survived in larger numbers. The phytoseiids seem well adjusted to the environment of the unsprayed orchards and to contribute to the low phytophagous mite populations in those orchards. In the sprayed orchards, the phytoseiids failed to control the high populations of phytophagous mites which developed. No statistical differences were found in the efficiencies of the mite brushing machine and Berlese funnels in removing either phytoseiid or phytophagous mites from apple leaves.