The paper addresses the problem of systemic analysis of architectures for distributed systems providing synchronous delivery of streaming multimedia content with coordinated control signals in watch-together scenarios. The relevance of the study is determined by the fact that user experience depends simultaneously on media delivery characteristics and the control plane, while inconsistency between these layers leads to desynchronization, delayed reactions, and degradation of interactivity. The problem lies in the complexity of architectural decision-making, where functional requirements for stream synchronization and non-functional system requirements must be considered prior to selecting specific protocols and implementation approaches. Methods: a systems analysis approach is applied with decomposition into media and control planes; fundamental architecture classes are identified based on the organization of media delivery (pull-based and push-based). Results: a model for selecting an architectural class is proposed, based on the analysis of functional and non-functional requirements for stream synchronization, supplemented by risk analysis of the chosen class that may necessitate requirement revision. Conclusions: the results can be used as a methodological framework for early architectural selection in the design of synchronous VOD systems and low-latency real-time scenarios, reducing the likelihood of late-stage architectural changes.