GABIDS (Genetic Algorithm Based Intrusion Detection System) is a dynamic IDS to support the real-time rule generation to identify abnormal network behavior. The chromosomes for the GA are essentially produced from information connectivity in the network. The currently existing IDS typically detect the attack type on basis of traditional static methods like fingerprinting. These IDS may not be robust enough to handle a new attack-type. In this paper, the authors propose GABIDS framework which dynamically generates the new set of rules to handle the unknown abnormal network behavior at any instant and update the rule-database accordingly.