In this study, we quantitatively tested the hypotheses, All network communication can be encrypted strongly at negligible costs. We used Twofish as a representative encryption algorithm of the future and analyzed its software performance across a class of devices. We analyzed encryption performance against various network standards. From our experimental results we see that, as far as the network is concerned, among devices we considered, which included a PalmPilot, an encryption rate of at least 100 kbps could easily be sustained. However, software encryption is an order of magnitude costlier compared to network transmission with respect to the power consumed. The latency due to encryption is acceptable, for communications over most networks and devices. But for small devices, this could overhead could become significant.
From the analysis in the previous sections, the broad picture that emerges is that universal software encryption is, in its entirety not feasible as of today. However, for certain applications, because of their varied demands and requirements, encryption may not be an overhead at all. We found this to be true for audio and video applications. For applications that require real-time guarantees, there was no significant latency due to encryption, when communication was over the internet. This means that encryption is feasible for applications such as those experimented upon in this study, and application developers can blindly use strong encryption for data in their distributed applications.