Nowadays, industrial control systems are falling squarely under the Industry 4.0 dominant trends and technological challengers. In this, the problems associated with data, their processing and analysis happen to be particularly serious. Among all data issues, data flows, most commonly called traffic flows, cause a deep concern since they are crucial for the correct real-time system functioning, and the knowledge of their characteristics is decisive for the assessment of critical time performance measures such as delays in alarm presentation and control action transmission. In order to explore industrial control system traffic flows and to calculate time performance measures, a theory of deterministic queuing systems Network calculus is worth to be utilized. The article is devoted to industrial control system traffic flows and the Network calculus application to them. In particular, the approaches to estimating envelopes for traffic flows are considered. The concept of envelopes on finite intervals is introduced, and their linear model is expounded. The advantages of linear envelopes on finite intervals over classical minimal envelope are revealed, and the algorithm for their online calculation is analyzed. The basic features of typical traffic flows encountered in upper-level control systems of nuclear power plants are outlined, and the results of the investigation of their deterministic envelopes as well as statistical indicators are presented. Conclusions are drawn about the legitimacy and practical acceptability of using linear envelopes on finite intervals in the applications of industrial control systems.