Typical AAL solutions rely on integrating capabilities for health monitoring, fall detection, communication and social inclusion, supervised physical exercises, vocal interfaces, robotic platforms etc. Ensuring the safe function and quality of service with respect to various extra-functional requirements like timing and security of such AAL solutions is of highest importance. To facilitate analysis, latest system development platforms provide underlying infrastructures for model-driven design (e.g., via the dime{} tool), timing and resource-usage specification (e.g., via the REMES tool), security features (e.g., by employing SECube), and statistical model-checking techniques (e.g, via Plasma). In this paper, we discuss the challenges associated with analyzing complex AAL solutions, from relevant properties to semantic interoperability issues raised by employing various frameworks for modeling and analysis, and applicability to evolving architectures. We take as examples two of the prominent existing AAL architectures and our own prior experience.