Allostery is a long-range macromolecular mechanism of internal regulation, in which the binding of a ligand in an allosteric site induces distant conformational changes in a distant portion of the protein, modifying its activity. From the drug design point of view, this mechanism can be exploited to achieve important therapeutic effects, since ligands able to bind allosteric sites may be designed to regulate target proteins. Computational tools are a valid support in this sense, since they allow the characterization of allosteric communications within proteins, which are essential to design modulator ligands. While considering long-range interactions in macromolecules, the principal drug design tool available to researcher is molecular dynamics, and related applications, since it allows the evaluation of conformational changes of a protein bound to a ligand. In particular, all-atoms molecular dynamics is suitable to verify the internal mechanisms that orchestrate allosteric communications, in order to identify key residues and internal pathways that modify the protein behaviour. The problem is that these techniques are heavily time-consuming and computationally intensive, thus high performance computing systems, including parallel computing and GPU-accelerated computations, are necessary to achieve results in a reasonable time. In this review, we will discuss how it is possible to exploit in silico approaches to characterize allosteric modulations and long-range interactions within proteins, describing the case study of the Heat Shock Proteins, a class of chaperons regulated by stress conditions, which is particularly important since it is involved in many cancers and neurodegenerative diseases.