A heuristic algorithm for minimizing server maintenance time and vulnerability surface on data centers

• Ano de defesa: 2020
• Autor(a) principal: Souza, Paulo Silas Severo de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul; Escola Politécnica; Brasil; PUCRS; Programa de Pós-Graduação em Ciência da Computação
• Palavras-chave em Português: Computação em nuvem; Manutenção de servidores; Virtualização; Segurança; Cloud computing; Server maintenance; Virtualization; Security; CIENCIA DA COMPUTACAO::TEORIA DA COMPUTACAO
• Link de acesso: http://tede2.pucrs.br/tede2/handle/tede/9522

Cost reduction and enhanced scalability boosted the adoption of cloud computing by multi-sized companies. To maintain the promised features, cloud operators perform several maintenance activities that range from removing dust accumulation to applying security patches on servers against vulnerabilities. The latter usually requires server update as soon as possible, as each instant servers need to wait for an update can indicate an opportunity for attackers to breach customers’ applications. Current solutions employ different approaches to minimize maintenance duration. However, they neglect the amount of time servers stay exposed to attacks. In this study, we first argue that only reducing maintenance time does not necessarily guarantee the efficiency of maintenance strategies on critical security patching scenarios, wherein safeguarding servers as soon as possible is the priority. Therefore, we propose a new metric called Vulnerability Surface, which aids operators in assessing the efficiency of maintenance strategies on critical security patching scenarios. Then, we present a heuristic algorithm that performs maintenance decisions to minimize the amount of time cloud servers remain exposed to attacks while reducing the amount of time necessary for performing server maintenance. We conducted a set of experiments against well-known strategies, and the results showed that the proposed solution achieves superior results regarding vulnerability surface, maintenance time, and the number of migrations per virtual machine.