The paper is about the clustering on large numeric data sets using hierarchical method. In this BIRCH approach is used, to reduce the amount of data, for this a hierarchical clustering method was applied to pre-process the dataset. Now a day’s web information plays a prominent role in the web technology, large amount of data is consumed to communicate, but some with intruders there is loss of data or may changes occur in the interaction, so to recognize intruders they detect to build an intrusion detection system for this a hierarchical approach is used to classify network traffic data accurately. Hierarchical clustering is performed By taking network as an example. The clustering method could produce high quality dataset with far less instances that sufficiently represent all of the instances in the original dataset.