Quantifying effect of network latency and clock drift on time-driven key sequencing

Abstract

Time-driven key sequencing (TKS) is a key management technique that synchronizes the session key used by a set of communicating principals based on time of day. This relatively low cost method of session key synchronization has been used in specialized distributed systems with low-end communicating devices where sessions are sparse and each session spans a short time period comprising a small number of messages. In this paper, we describe how TKS may be useful in several scenarios involving high speed computer networks. More importantly, we present a performance model of TKS and conduct a detailed analysis to determine the impact of clock drift and network latency on the required key refresh rate. We give the exact conditions for determining the range of adequate key refresh rates, and demonstrate that the derived conditions are sufficient to ensure that data are both protected and deliverable. Interestingly, these conditions may be used to obtain a key refresh rate that can tolerate a maximum amount of clock drift after other parameters in the system are fixed.