Disclosed herein are systems and methods for multi-level classification of data traffic flows based on information in a first packet for a data traffic flow. In exemplary embodiments of the present disclosure, a key can be generated from intercepted DNS data to track data traffic flows by application names and source and destination IP addresses. Based on these keys, patterns can be discerned to infer data traffic information based on only the information in a first packet, such as source and destination IP address. The determined patterns can be used to predict classifications of future traffic flows with similar key information. In this way, data traffic flows can be classified and steered in a network based on limited information available in a first packet.