The aim of the work is to study robustness shape measures and build on their basis a display space of possible implementations of the Burr type XII distribution, which made is possible to modeling the reliability of processes in all areas of human activity. The correct choice of the implementation shape of the reliability model determines the reliability of the forecast for computing networks and devices of exchange information that are used in modern high-tech production processes. In this paper robust measures of distribution shape are proposed to use when new methods for choosing a distribution model is implemented. Since robust measures are not containing any restrictions on the shape parameters these measures make is possible to be extended the study to heavy-tailed distribution. On the basis of the Shannon information measure, formulas for the entropy measures of the Burr type XII distribution are obtained and a robust entropy coefficient is proposed as a new robust shape measure. The main result of the study is a three-dimensional space of robust measures that is containing all possible implementations of the Burr XII type distribution.