Over the past few decades, two complementary concepts, smart environments (SEs) and human activity recognition (HAR), have experienced significant growth. Recent developments in HAR provide SEs with greater intelligence in identifying specific situations such as heart attacks, falls and spatial location, among others. This highlights the need for new architectures that enable seamless integration of devices and HAR techniques for structured data sharing and processing. This will enable comprehensive analysis by bringing together all the information available in SEs to generate context during data analysis for identifying more complex situations. This first part of the work presents the design and implementation of a highly scalable architecture, called Home-Monitor, that enables the integration of HAR components with heterogeneous data acquisition and output devices. The architecture allows for highly distributed execution that is robust enough to allow for subsystems to be removed or added at runtime. It also facilitates rapid prototyping for data collection, analysis and output using traditional web technologies such as HTTP-REST services and JSON messages.