This research analyzes and compares the digital carbon footprint of air quality data processing (PM2.5 and CO2) using two computing infrastructures: Virtual Private Server (VPS) and Cloud Computing. Moving beyond simple technical comparison, this study establishes measurable causal relationships between infrastructure choices and environmental impact through a systematic methodology. Applying the Sustainable Web Design Model (SWDM) and validating results through the Website Carbon Calculator and Green Web Foundation API, the study quantifies both operational and embodied emissions. The results show that the optimized Cloud Computing infrastructure generates 0.30gCO2 per user visit (Rating E), while the VPS in Indonesia produces 0.92gCO2 per visit (Rating F), representing a 67.4% reduction in emissions. This significant difference is primarily attributed to three factors: (1) server location and energy source, with Cloud infrastructure leveraging lower-carbon regions via Content Delivery Networks; (2) microservices architecture enabling efficient resource utilization; and (3) Docker containerization optimization, which reduced the image size by 98.85% (from 4.3GB to 49.45MB), directly contributing to emission reduction and a 64.3% improvement in load time. These findings highlight the critical impact of server location, architectural design, and infrastructure optimization on digital carbon emissions—particularly in regions like Indonesia, where electricity is predominantly fossil fuel-based. The study provides a foundation for policy recommendations on migrating to environmentally friendly computing infrastructures, especially for real-time IoT-based environmental data processing.