605 research outputs found
The ATLAS experiment on-line monitoring and filtering as an example of real-time application System monitoringu i filtracji eksperymentu ATLAS jako przykład aplikacji czasu rzeczywistego /
No full textTyt. z nagłówka.Pozostali autorzy artykułu: Tadeusz Szymocha, Włodzimierz Funika, Jacek Kitowski, Renata Słota, Kazimierz Bałos Łukasz Dutka, Krzysztof Guzy, Tomir Kryza, Jan Pieczykolan.Bibliografia s.86.Dostępny również w formie drukowanej.ABSTRACT: The ATLAS detector, recording LHC particles' interactions, produces events with rate of 40 MHz and size of 1.6 MB. The processes with new and interesting physics phenomena are very rare, thus an efficient on-line filtering system (trigger) is necessary. The asynchronous part of that system relays on few thousands of computing nodes running the filtering software. Applying refined filtering criteria results in increase of processing times what may lead to lack of processing resources installed on CERN site. We propose extension to this part of the system based on submission of the real-time filtering tasks into the Grid. STRESZCZENIE: Detektor ATLAS, rejestrujący zderzenia protonów rozpędzanych w zderzaczu LHC, będzie generował przypadki o rozmiarze 1.6MB z częstotliwością 40MHz. Aby wyselekcjonować bardzo rzadko występujące przypadki z interesującymi oddziaływaniami fizycznymi, konieczne będzie zastosowanie wydajnego systemu filtracji (trigger). Część asynchroniczna takiego systemu wykorzystuje kilka tysięcy komputerów, na których wykonywane jest oprogramowanie filtrujące. Zwiększenie selektywności systemu wymaga zwiększenia czasu procesowania, co może doprowadzić do wyczerpania zasobów komputerowych zainstalowanych w CERN-ie. Proponujemy rozszerzenie tej części systemu poprzez umożliwienie wykonywania oprogramowania filtrującego w czasie rzeczywistym na komputerach w środowisku gridowym. KEYWORDS: high energy physics, real-time procesing, trigger system, remote farms. SŁOWA KLUCZOWE: fizyka wysokich energii, przetwarzanie w czasie rzeczywistym, system filtracji, zdalne farmy
Unified metadata management in large distributed computing infrastructures
No full textPromotor: Jacek Kitowski, Renata Słota.Recenzent: Marian Wysocki, Bogdan Wiszniewski.Niepublikowana praca doktorska.Tyt. z ekranu tyt.Praca doktorska. University of Science and Technology AGH in Kraków. Faculty of Computer Science, Electronics and Telecommunications. Department of Computer Science, 2014.Zawiera bibliogr.Dostępna również w wersji drukowanej.Tryb dostępu: Internet.Grid computing, services and cloud infrastructures, evolution of Grid computing, virtual organizations, models of virtual organizations, contracts in virtual organizations, virtual organizations in grid computing, existing Grid environments, globus toolkit, UNICORE, gLite, QosCosGrid, most important standards and technologies in Grid computing, OGSA, WS-resource framework, security, information and monitoring services, interoperability between Grid frameworks, service oriented architectures, SOAP, WSDL, UDDI and WS-I, BPEL and WS-CDL, REST, enterprise service bus, P2P computing, emergence of cloud computing paradigm, metadata and ontologies, metadata in general, metadata in computing systems, Simple Knowledge Organization System SKOS, GLUE Schema, Common Information Model, knowledge representation formalisms, description logics, OWL Web Ontology Language, OWL 2, Semantic Web Knowledge Bases, unification of metadata in distributed computing infrastructures, through ontologies, motivation for metadata unification, description logics as metadata representation formalism, ontology separation scheme, Generic Grid Ontologies, ontology of resources, ontology of data, ontology of services, example use case, integration of common information model metadata scheme to OWL, managing semantic metadata with Grid Organizational Memory, requirements for a Grid semantic metadata repository, architecture of Grid Organizational Memory, GOM architecture, GOM engine, engine manager, deployer, GOMAdmin, proxy, protege plugin GOMTab, usage scenarios, performance evaluation, GOM Event System and knowledge evolution, GOM Peer-to-peer distribution model, interfacing Legacy Metadata Systems, LDAP integration to semantic knowledge base, semantic framework for virtual organizations, semantic approach to virtual organizations, framework architecture, support for contract negotiation, contract negotiation model, contract negotiation interfac
Interactive cloud data farming environment for military mission planning support
No full textTyt. z nagłówka.Pozostali autorzy artykułu: Dariusz Kryza, Michal Wrzeszcz, Lukasz Dutka, Jacek Kitowski.Bibliogr. s. 98-100.In a modern globalised world, military and peace keeping forces often face situations which require very subtle and well planned operations taking into account cultural and social aspects of a given region and its population as well as dynamic psychological awareness related to recent events which can have impact on the attitude of the civilians. The goal of the EUSAS project is to develop a prototype of a system enabling mission planning support and training capabilities for soldiers and police forces dealing with asymmetric threat situations, such as crowd control in urban territory. In this paper, we discuss the data-farming infrastructure developed for this project, allowing generation of large amount of data from agent based simulations for further analysis allowing soldier training and evaluation of possible outcomes of different rules of engagement.Dostępny również w formie drukowanej.KEYWORDS: data farming, cloud, virtualisation, web 2.0, mission planning suppor
Self-healing highly available monitoring system for distributed environment
No full textPromotor: Jacek Kitowski, Włodzimierz Funika.Recenzent: Marian Wysocki, Stanisław Kozielski.Niepublikowana praca doktorska.Tyt. z ekranu tyt.Praca doktorska. AGH University of Science and Technology. Faculty of Computer Science, Electronics and Telecommunications. Department of Computer Science, 2014.Zawiera bibliogr.Dostępna również w wersji drukowanej.Tryb dostępu: Internet.High availability, self-healing, agent based systems, reliability, technology background, discovery and communication, Service Location Protocol SLP, Simple Service Discovery protocol SSDP, JINI, JXTA, JGroups, election algorithms, Bully Algorithm, Ring Algorithm , Santoro Rotem Algorithm, other algorithms, overview of existing monitoring systems, Ganglia, AutoPilot, Gemini, Aksum, JavaPSL, OCM-G / G-PM, J-OCM, JXM, SemMon, Dynamic Monitoring Framework, self-healing and self-adaptive Software, PANACEA, MUSIC, GRAVA, use cases and requirements, introduction, use cases, user use cases, advanced user administrator use cases, application programmer use cases, system requirements definition, functional requirements, non-functional requirements, functional requirements specification, non-functional requirements specification, architecture and design, introduction, distributed system, multi-agent based system MAS, self-healing, distributed monitoring system architecture, AgeMon architecture, agent, agent name and collision detection, monitoring result, monitoring source and measurable capability, messaging, agent communication Layer, discovery finding agents, discovery based on the multicast, discovery based on the Gossip Server , tunneling, discovery, mixed solution, roles, GUI Role, monitoring manager, visualization and visualization manager, rules manager component, monitoring role, handling feedbacks enabling the healing of the monitored system, persistence role, rule role, command line interface role, system implementation, introduction, agent, agent group, role implementation, monitoring role, GUI role, AgentGraph, monitoring component, rule component, persistence component, visualisation component, rule role , persistence role, CLI role, system-wide architectural concepts, service interface and start up/shutting down procedure, foundation services, dependency injection, communication in the AgeMon System, transport layer , abstract agent layer, abstract monitoring layer, message types, high availability and self-healing, high availability and self-healing in monitoring system, automatic discovery, reliable transport protocols, network failures tolerance, absence of single point of failure, roles redundancy, substitute agents – failover, cooperative mode – rules, advanced rules, self-healing, high availability/self-healing strategies, healing of the monitored system, transparent healing, application aware healing , testing, system deployment, performance tests, scenario 1, scenario 2, scenario 3, performance of the low-level communication layer, latency in the system, decision time, persistence time, network latencies, latency summary, self-healing tests / high availability tests, substitute persistence agent, restart monitoring agent, Healing SUM, soak test, reliability, code quality, code statistics , violations, duplications, cyclomatic complexity, LCOM4, code quality, novel concepts introduced, future development, enlarge agent autonomy, reasoning from the historical results, advanced data analysis, learning, advanced metrics transformations, rules and action enhacements, complex event processing, advanced fault detection in monitoring system, GUI Enhancements – support for different types of charts, non-relational database
SLA-oriented semi-automatic management of data storage and applications in distributed environments Semiautomatyczne zarządzanie aplikacjami oraz składowaniem danych w środowiskach rozproszonych z uwzględnieniem parametrów SLA /
No full textTyt. z nagłówka.Pozostali autorzy art.: Włodzimierz Funika, Renata Słota, Jacek Kitkowski.References. p.49-50.Dostępny również w formie drukowanej.STRESZCZENIE: Artykuł opisuje semiautomatyczny szkielet aplikacyjny służący do wsparcia procesu wdrażania aplikacji oraz składowania dużych ilości danych w środowiskach rozproszonych z uwzględnieniem parametrów jakościowych. Projekt PL-Grid ma na celu wsparcie polskiej nauki w celu umożliwienia naukowcom przeprowadzania złożonych eksperymentów typu in-silico na skalę wiekszą niż dotychczas. W artykule zostały opisane wyzwania związane ze strategiami zarządzania wielkimi ilościami danych, zdefiniowane w fazie analizowania wymagań użytkowników projektu PL-Grid. Zostały również opisane proponowane rozwiązania omawianych problemów, opis przykładowych scenariuszy użycia oraz aktualny stan prac implementacyjnych i rezultaty przeprowadzonych testów. SŁOWA KLUCZOWE: składowanie danych, zarządzanie danymi, środowiska rozproszone. ABSTRACT: In this paper we describe a semi-automatic programming framework for supporting users with managing the deployment of distributed applications along with storing large amounts of data in order to maintain Quality of Service in highly dynamic and distributed environments, e.g., Grid. The Polish national PL-GRID project aims to provide Polish science with both hardware and software infrastructures which will allow scientists to perform complex simulations and in-silico experiments on a scale greater than ever before. We highlight the issues and challenges related to data storage strategies that arise at the analysis stage of user requirements coming from different areas of science. Next we present a solution to the discussed issues along with a description of sample usage scenarios. At the end we provide remarks on the current status of the implementation work and some results from the tests performed. KEYWORDS: data storage, application management, distributed environment
Management methods in SLA-aware distributed storage systems
No full textTyt. z nagłówka.Pozostali autorzy artykułu: Renata Słota, Danilo Lakovic, Paweł Winiarczyk, Marek Pogoda, Jacek Kitowski.Bibliogr. s. 43.Traditional data storage systems provide access to user’s data on the “best effort” basis. While this paradigm is sufficient in many use cases it becomes an obstacle for applications with Quality of Service (QoS) constraints. Service Level Agreement (SLA) is a part of the contract agreed between the service provider and the client and contains a set of well defined QoS requirements regarding the provided service and the penalties applied in case of violations. In the paper we propose a set of SLA parameters and QoS metrics relevant to data storage processes and the management methods necessary for avoiding SLA violations. A key assumption in the proposed approach is that the underlying distributed storage system does not provide functionality for resource or bandwidth reservation for a given client request.Dostępny również w formie drukowanej.KEYWORDS: service level agreements, mass storage systems, quality of service
Benchmarking high performance architectures with natural language processing algorithms Benchmarking architektur wysokiej wydajności algorytmami przetwarzania języka naturalnego/
No full textTyt. z nagłówka.Bibliogr. s. 30-31.Algorytmy przetwarzania języka naturalnego mają duże zapotrzebowanie na zasoby komputerowe, szczególnie gdy wymagane jest dostosowanie algorytmu do języka fleksyjnego jakim jest np. język polski. Artykuł przedstawia wymagania czasowe i pamięciowe algorytmów tagowania częściami mowy oraz algorytmów klasteryzacji zastosowanych do dwóch korpusów języka polskiego. Dokonano benchmarkingu algorytmów na trzech platformach wysokiej wydajności reprezentujących różne architektury. Dodatkowo porównano wersję sekwencyjną oraz implementacje OpenMP algorytmów klasteryzacji.Natural Language Processing algorithms are resource demanding, especially when tuning to inflective language like Polish is needed. The paper presents time and memory requirements of part of speech tagging and clustering algorithms applied to two corpora of the Polish language. The algorithms are benchmarked on three high performance platforms of different architectures. Additionally sequential versions and OpenMP implementations of clustering algorithms were compared.Dostępny również w formie drukowanej.SŁOWA KLUCZOWE: benchmarking, tagowanie częściami mowy, klasteryzacja dokumentów, przetwarzanie języka naturalnego. KEYWORDS: benchmarking, part-of-speech tagging, document clustering, natural language processing, high performance architectures
Mobile social networks for live meetings
No full textTyt. z nagłówka.Bibliogr. s. 99-100.In this article, we present an idea of combining social networking websites and modern mobile devices abilities to transfer social networking activity to a higher level. Nowadays, these devices and websites are used to offer ability of remote communication (phone calls, message exchange etc.), which potentially can be used to notify people about meetings in the real world. Since the current social network models do not provide enough information for such notification (social networking websites are examples of social networks) a new social network model that will be suitable for the above mentioned application is proposed and a new social platform that base on mobile devices is introduced. This platform can notify users when their friends are nearby. The paper presents the model and the simulation that verifies the approach.Dostępny również w formie drukowanej.KEYWORDS: social networks, social networking websites, mobile devices
Massively self-scalable platform for data farming
No full textPromotor: Jacek Kitowski.Recenzent: Krzysztof Cetnarowicz, Bogdan Wiszniewski.Niepublikowana praca doktorska.Tyt. z ekranu tyt.Praca doktorska. Akademia Górniczo-Hutnicza im. Stanisława Staszica (Kraków). Wydział Informatyki, Elektroniki i Telekomunikacji. Katedra Informatyki, 2014.Zawiera bibliogr.Dostępna również w wersji drukowanej.Tryb dostępu: Internet.Data farming, self-scalable software, heterogeneous computational infrastructures, problem description, thesis statement and research objectives, note on participation in European research projects, background survey, data farming systems, OldMcData, JWARS, SWAGES, DIRAC, self-scalable systems, staged event-driven architecture, gigaspaces eXtreme application platform, teradata database, apache hadoop, computational environments, grid computing, cloud computing, massively self-scalable platform, concept, architecture, development methodology for a data farming platform, platform use cases, data farming use cases, platform management use cases, massive self-scalability requirement, concept of self-scalable services, self-scalable services in the data farming platform, problem of scalability, motivation for scalability, scalability metrics, common scaling strategies and potential bottlenecks, scaling rule definition, scalability in scalarm platform, scalarm implementation details, platform overview, scalarm services, experiment manager, storage manager, simulation manager, information manager, node manager, monitoring, scalability manager, load balancer, cache, architectural elements supporting scalability, automatic scalability management, implementation of essential use cases, creating a data farming experiment use case, simulation execution use case, extending an experiment use case, experimental evaluation, evaluation objectives, evaluation of massive scalability, testing scenario, testing environment, scalability evaluation results, self-scalability evaluation, self-scalability test scaling rules disabled, self-scalability test with scaling rules for experiment manager, data farming utilization in training of security force
A toolkit for storage QoS provisioning for data-intensive applications
No full textTyt. z nagłówka.Pozostali autorzy artykułu: Dariusz Król, Kornel Skałkowski, Michał Orzechowski, Darin Nikolow, Bartosz Kryza, Michał Wrzeszcz, Jacek Kitowski.Bibliogr. s. 71-72.This paper describes a programming toolkit developed in the PL-Grid project, named QStorMan, which supports storage QoS provisioning for data-intensive applications in distributed environments. QStorMan exploits knowledgeoriented methods for matching storage resources to non-functional requirements, which are defined for a data-intensive application. In order to support various usage scenarios, QStorMan provides two interfaces, such as programming libraries or a web portal. The interfaces allow to define the requirements either directly in an application source code or by using an intuitive graphical interface. The first way provides finer granularity, e.g., each portion of data processed by an application can define a different set of requirements. The second method is aimed at legacy applications support, which source code can not be modified. The toolkit has been evaluated using synthetic benchmarks and the production infrastructure of PL-Grid, in particular its storage infrastructure, which utilizes the Lustre file system.Dostępny również w formie drukowanej.KEYWORDS: data-intensive application, storage management, QoS, Grid
- …
