Next: Conclusions
Up: Specifics of the solution
Previous: Experimental Framework
If preliminary analysis show that universal encryption is not
possible, we propose to identify and characterize a class of
applications and environment where the encryption is feasible or not
feasible at low costs. The graphs obtained by the method described
previously, will help us in that task. For example, if a PDA lies
below the IrDA line in the speed of software encryption graph, clearly
we cannot run any secure application designed for IrDA and involving
the PDA.
We propose to use the experimental framework described above, to
analyze typical applications for performance evaluation. Applications
in a c0pe environment where encryption performance is critical can
broadly be categorized as,
- 1.
- Constant Bit Rate (CBR) with soft real time requirements, eg audio.
- 2.
- Variable Bit Rate (VBR) with soft real time requirements, eg video.
- 3.
- Hard real time applications, eg control systems, eg fire safety alarms
We will study representative applications in those categories for
evaluating the relevant costs of encryption. For example, a audio
application demands that a constant rate of encryption be sustained.
A video application may involve bursts of traffic and hence a device
should be able to accommodate such peak rates. Finally, control
systems may demand that latency of encryption operations be within a
defined upper bound.
Next: Conclusions
Up: Specifics of the solution
Previous: Experimental Framework
Ramadass Nagarajan
2000-03-21