Cloud query manager: uso de web sem?ntica para evitar o problema de aprisionamento em IaaS

• Main Author: Souza, Arthur Emanoel C?ssio da Silva e
• Other Authors: 03038841420
• Language: Portuguese
• Published: Universidade Federal do Rio Grande do Norte 2016
• Subjects: CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::SISTEMAS DE COMPUTACAO; Computa??o em nuvem; Modelo h?brido; SPARQL
• Online Access: http://repositorio.ufrn.br/handle/123456789/20154

Submitted by Automa??o e Estat?stica ([e-mail]) on 2016-03-31T22:29:13Z No. of bitstreams: 1 ArthurEmanoelCassioDaSilvaESouza_DISSERT.pdf: 1480067 bytes, checksum: 15c2cf114fae56a6c186a5417cecb57f (MD5) === Approved for entry into archive by Arlan Eloi Leite Silva ([e-mail]) on 2016-04-04T19:16:06Z (GMT) No. of bitstreams: 1 ArthurEmanoelCassioDaSilvaESouza_DISSERT.pdf: 1480067 bytes, checksum: 15c2cf114fae56a6c186a5417cecb57f (MD5) === Made available in DSpace on 2016-04-04T19:16:06Z (GMT). No. of bitstreams: 1 ArthurEmanoelCassioDaSilvaESouza_DISSERT.pdf: 1480067 bytes, checksum: 15c2cf114fae56a6c186a5417cecb57f (MD5) Previous issue date: 2015-01-28 === A computa??o em nuvem pode ser definida como um modelo computacional distribu?do atrav?s do qual s?o compartilhados recursos (hardware, armazenamento, plataformas de desenvolvimento e comunica??o) na forma de servi?os pagos e acess?veis com um esfor?o m?nimo de gest?o e intera??o. Um grande benef?cio desse modelo ? possibilitar a utiliza??o de variados provedores (i.e. uma arquitetura multiplataforma) para compor um conjunto de servi?os no intuito de obter uma configura??o ?tima em desempenho e custo. Entretanto, a utiliza??o multiplataforma ? impossibilitada pelo problema de cloud lock-in. O cloud lock-in ? a depend?ncia entre uma aplica??o e uma plataforma de nuvem. Ele ? comumente endere?ado por tr?s estrat?gias: (i) uso de camada intermedi?ria que se interp?e aos consumidores de servi?os de nuvem e o provedor, (ii) uso de interfaces padronizadas para acesso a nuvem, ou ainda (iii) uso de modelos com especifica??es abertas. Este trabalho define uma abordagem para avaliar estas estrat?gias. Tal abordagem foi executada e observou-se que apesar dos avan?os conseguidos por estas estrat?gias, nenhuma delas resolve efetivamente o problema do cloud lock-in. Neste sentido, esse trabalho prop?e o uso de Web Sem?ntica para evitar o cloud lock-in, onde modelos em RDF s?o usados para especificar os recursos de uma nuvem, que s?o gerenciados via consultas SPARQL. Nessa dire??o, esse trabalho: (i) apresenta um modelo avaliativo que quantifica o problema do cloud lock-in, (ii) avalia o cloud lock-in entre tr?s solu??es multi-cloud e tr?s plataformas de nuvem, (iii) prop?e utilizar RDF e SPARQL na ger?ncia de recursos de nuvem, (iv) apresenta o Cloud Query Manager (CQM), um servidor SPARQL que implementa a proposta, e (v) compara tr?s solu??es multi-cloud em rela??o ao CQM quanto ao tempo de resposta e a efetividade na resolu??o do cloud lock-in. === Cloud computing can be defined as a distributed computational model by through resources (hardware, storage, development platforms and communication) are shared, as paid services accessible with minimal management effort and interaction. A great benefit of this model is to enable the use of various providers (e.g a multi-cloud architecture) to compose a set of services in order to obtain an optimal configuration for performance and cost. However, the multi-cloud use is precluded by the problem of cloud lock-in. The cloud lock-in is the dependency between an application and a cloud platform. It is commonly addressed by three strategies: (i) use of intermediate layer that stands to consumers of cloud services and the provider, (ii) use of standardized interfaces to access the cloud, or (iii) use of models with open specifications. This paper outlines an approach to evaluate these strategies. This approach was performed and it was found that despite the advances made by these strategies, none of them actually solves the problem of lock-in cloud. In this sense, this work proposes the use of Semantic Web to avoid cloud lock-in, where RDF models are used to specify the features of a cloud, which are managed by SPARQL queries. In this direction, this work: (i) presents an evaluation model that quantifies the problem of cloud lock-in, (ii) evaluates the cloud lock-in from three multi-cloud solutions and three cloud platforms, (iii) proposes using RDF and SPARQL on management of cloud resources, (iv) presents the cloud Query Manager (CQM), an SPARQL server that implements the proposal, and (v) comparing three multi-cloud solutions in relation to CQM on the response time and the effectiveness in the resolution of cloud lock-in.