Preventive method based on ranking and voting scheme for intrusion detection in webservice

• Autor(a) principal: Danta, Ligia Maria da Silva
• Data de Publicação: 2021
• Outros Autores: Guelfi, Adilson Eduardo, da Silva, Anderson Aparecido Alves, de Azevedo, Marcelo Teixeira, Kofuji, Sergio Takeo
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: https://doi.org/10.34627/rcc.v16i0.263

The Price and Deadline Calculator (CPP) is a gratuitous webservice available on the Internet for freight consultations that receive between 8 and 28 thousand of daily queries. CPP constantly receives attacks in the form of queries that cause unavailability and slowness in the system. The current criteria used to solve this problem are not based on traffic volume, because it is costly and may eventually block valid customers. This work aims to propose a method of preventive detection of attacks against availability, based on a rankings/weights scheme that improves the accuracy and response time of detecting suspicious connections in a webservice. In the experiment, a voting scheme is created, generated from network traffic, where standard-based statistics and predefined rules are used to compose a ranking of suspicion for each query.