Apart from great convenience derived, computer networks have grown into the best intrusion channel for those who aspire to. This paper exploits a conglomeration of Artificial Immune System and Finite State Machine to build highly robust intrusion detection algorithm in order to reinforce information security. AIS is widely adopted in various disciplines, as it hold the advantage of achieving better diverity inherently than genetic algorithm does. Aiming to better identification in non-consecutive packet attacks and novel anomaly, our immunization algorithm takes inspiration from Wright's island model, gene migration, and heterosis of population genetics, producing effectual outcomes. The results shows that our algorithm is able to capture more novel attack patterns while maintaining low false positive rate.