Biomedical time-series analysis deals with detection, classification and prediction of subjects' states and disorders. In this paper, we present requirements and scenarios of use for a novel web platform designed to analyze multivariate heterogeneous biomedical time-series. The scenarios of use are described with the corresponding UML Use Case Diagrams. We also discuss some architectural and technological issues, including parallelization, visualization and feature extraction from biomedical time-series. The goal of this paper is to present what we currently consider as the best approach for design of such a system, which may also be beneficial for similar biomedical software systems. The paper is focused on design and architectural considerations only, as implementation of the complex system has only just begun.