This paper describe a Stream Classification Adaptive Model (SCAM) technique for securing massive sensing information streams that meets multiple levels of confidentiality and integrity. This SCAM technique includes two vital concepts: common shared keys that area unit initialized and updated by D-SM while not requiring retransmission and a seam-less key stimulant method while not break off the data-stream encryption/decryption. Moreover, a replacement theme is planned to secure a big data sensor protocol through the employment of multiple unidirectional hash chains. The theme is shown to be lower in machine, power utilization. Also, communication prices area unit nevertheless still ready to secure Big Data Sensor communication model.