The Wireless Sensor Networks (WSNs) are composed of a huge number of sensor nodes. The large number of nodes may lead to a huge amount of data in the network, causing the network to degrade performance and shorten its lifetime. The data aggregation techniques may be a solution to remane the redundancy of data and thereby reduce the number of packets transmitted in the network. Nevertheless, the data aggregation causes some security vulnerabilities: impersonation, denying having received data, dropping packets deliberately, alteration of the sensored readings and aggregation results... In this report, we compare the performances get by two secure aggregation solutions: the Secure Aggregation for Wireless Networks (Hu et al. protocol) and the Secure Aggregation Protocol for Cluster-Based Wireless Sensor Network (SAPC). The analysis shows that Hu et al. protocol gives better performance than SAPC in terms of number of exchanged messages but it is vulnerable to simple attacks. This report describes both protocols and their performances and propose a new protocol based on binary trees and delayed aggregation result checking to improve the performance of SAPC