Abstract
The minimal 6TiSCH configuration provably suffers from large initial synchronization times, which results in an almost linear increase in the node energy consumption. One way to reduce the initial synchronization time is the use of an optimal scan period. Although the optimal scan period value has already been found in the literature, this value only takes into account the case where the joining node knows the number of channels used by the network. In this paper, we conduct an extensive analysis on the initial synchronization in the minimal 6TiSCH configuration when a joining node - that is a node desiring to join a 6TiSCH network - does not know the network channels, which is a scenario closer to the reality. Our analysis, which includes a mathematical model, large-scale simulations, and experiments in a testbed, demonstrates that when a joining node does not know the network channels, we can lower the maximum possible average synchronization time, that is, we can guarantee a lower upper-bound in the average synchronization time by setting a sub-optimal value. This value can be calculated by the joining node without knowing the network channels. By using a sub-optimal scan period, we can achieve an up to 82.5% and 89% lower upper-bound in the average initial synchronization time compared to the default scan periods of the protocol implementations in Contiki-NG and OpenWSN, respectively.