In this section, the performance of FTPR protocol was compared with the GTMS  protocol in terms of communication overhead and memory consumption.
In FTPR, when a node from ith cluster wants to communicate with the sink through its CH, the total number of acknowledgments generated for trust computation was 2(h - 1).
FTPR broadcasts one recommendation request and receives recommendation only from a set of neighbors and let a be the number of received recommendations.
Communication overhead of GTMS protocol increases with cluster size as shown in Table 4, whereas the communication overhead of FTPR protocol was same throughout as it was not dependent on cluster size.
As a result, the communication overhead of GTMS-18 is 17.9 percent higher than our proposed FTPR protocol.
But in FTPR protocol, all the nodes used similar low power transmitters and communicated with the sink using a multihop link.
In FTPR, CMs and CHs maintained a transaction table to monitor and store the trust level of their neighbors.
The memory consumption for FTPR and GTMS protocols was plotted against the number of neighboring nodes and setting the size of the observation window n = 4 as shown in Table 7.
It was found that the memory consumption in FTPR protocol is 19.9 percent lower than the GTMS protocol.
In this paper, we proposed FTPR protocol to effectively thwart black hole attack, on-off attack, conflicting behavior attack, and bad-mouthing attack.
Theoretical and simulation results of FTPR protocol demonstrate higher packet delivery ratio, lower end-to-end delay, higher network lifetime, and lower memory consumption than the traditional and existing trust-based routing schemes.