When solving fundamental scientific problems using the BOINC infrastructure, a number of systemic effects arise. These effects are specific to the BOINC system and cannot be observed when using portable batch systems. This is due to the different operating mechanisms of voluntary computing systems and batch systems. An important distinguishing feature is the behavior of computing nodes when there are no jobs in the system. The paper examines three scenarios that are often encountered in practice for the operation of a computing infrastructure: temporary deactivation of a project, dynamics of request execution after deactivation, and activation of a project after long-term inactivity. Consideration of each aspect is based on the analysis of the principle of Binary Random Exponential Backoff. The research results will allow us to more deeply understand the patterns that arise when working with BOINC. It will be useful both in the field of practical calculations and for further probabilistic research.