Today's world i. e. either private or public (government) sector totally influenced by Internet and networking for their business, entertainment purpose. But black side of an internet cannot be ignored . Internet makes the door open to the intruders and hackers. Any successful attempt made by intruders in capturing data causes a big loss of confidential data or digital money. This forces organizations to adopt serious security policy. It may include either use of an encryption methodology or firewall . Now a days as preventive security organization may use an Intrusion detection system. An Intrusion detection system classifies incoming data as normal data and attack data and if attack pattern is recognized it gives alarm to the network administrator . There are various algorithms and methods are proposed for IDS. In this paper we proposed the importance of ANN in developing NIDS for detecting denial of service attacks. The proposed ANN based NIDS will make classification of input packets into 4 categories DoS, U2R,R2L, PROBING in the first step . In the second step NIDS system can be enhanced to detect the DoS attack as Smruf, Teardrop, Neptune, Land, Pod, Back. This developed NIDS can minimize false positive and false negative rate by increasing number of hidden layers in the construction of ANN.

Artificial Neural Network (ANN) . Network Intrusion Detection System (NIDS) Denial of service (DoS).