Detecção de falhas e re-roteamento automático e transparente de pacotes para serviços de rede com redundância implementada

Abstract

Replication is frequently used to increase the availability of network services. Redundancy is the direct result of component or service replication. Replication, when properly implemented, can prevent clients (users or systems) to keep without access to services during temporary or permanent periods. The goal of this work is to propose a software solution that is capable of providing higher levels of availability in environments that have redundancy implemented in essential network services such as DNS, LDAP and SMTP. Clients who use the network services typically have static configurations or support only a single address, to one replica. Having this in mind, the goal of the work is to increase the availability rate, from the client perspective, by applying transparent and automated techniques, such as packet re-routing. To achieve this result, the idea is to use resources and tools available inside firewalls, which are important network componentes usually present in local networks. This work presents the solution called AvailableD. It automatically and transparently increases the availability of network services with implemented redundancy. To do so, it uses resources within GNU/Linux firewalls. The solution was designed with simplicity and ease of use in mind. It consists of three basic elements: a) plug-ins for services monitoring; b) packet re-routing mechanisms; c) resources for sending notifications to network administrators. The first step is to detect a failure of any service replica. The second step is to apply packet rerouting rules, so that the requests addressed to failed replicas are redirected to active replicas. Finally, failures in replicas are reported to network administrators. The AvailableD is implemented as an open source, extensible and adaptable tool. The adaptability is achieved through an architecture which is based on plug-ins that perform tasks such as services monitoring, re-routing rules management and notifications delivery. The solution has been tested and evaluated in a scenario similar to the NTIC’s, of UNIPAMPA, environment. The results show that the main goal of the work was reached through the transparent and automated increase of the availability rate for network services with implemented replication.