Cloudlets móveis : deslocamento de serviços para redução de latência

• Ano de defesa: 2016
• Autor(a) principal: Rezende, Rodrigo de Ávila
• Orientador(a): Salgueiro, Ricardo José Paiva de Britto
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Sergipe
• Programa de Pós-Graduação: Pós-Graduação em Ciência da Computação
• País: Brasil
• Palavras-chave em Português: Computação; Computação em nuvem; Computação móvel; Servidores da Web; Armazenamento de dados; Banco de dados; Cloudlet; Mobilidade; Docker; Gerência de banco de dados
• Palavras-chave em Inglês: Handoff
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://ri.ufs.br/handle/riufs/3366

Cloud Service have been usual to solve several computer issues, some companies like Amazon or Microsoft decide invest a big amount of money to sell computer resources as PaaA or IaaS. Subsequently, new devices, now portables, emerge as common and needs cloud computing to provide resources, creating a new structure known as mobile cloud computing. Even showing several models to its architecture in mobile cloud computing, natural behavior converge to total transparency to applications and users concerning to resource location, requiring that mobile devices maintains a permanent connection and an approach that needs a procces mirroring, knows as Cloudlets. These Cloudlets should provides required resources to guarantee a cofortable perception experience to users, giving an alternate option to latency problems, local processing and unavailable connections. To obtain this quality of perception, the Cloudlet should be closest to user as possible and follow him all the time during his travel. The usual model about Cloudlets consists on services running at virtual machine, that are costly what concerns moving and startup time. As alternative this work proposes a docker container based model, which will permit a smaller startup time less tranfer data about system between datacenters evolved. This alternative shows, at proposed scenarios, is possible reduce startup overhead and moving time, benefeting users perception and opening new perspectives to implement and optimizing.