zWang64/CloudApps4Rural

While efficiency metrics (e.g., performance and energy) are certainly critical to gauge cloud systems, we must start to think about the societal implications of the web systems we build. We are exploring intersectionality, equity, and fairness as first-order web system design metrics. Specifically, instead of continuing to build web systems that serve the top 1% of the world's population, we are interested in building cloud systems for rural communities, starting with rural African communities (we are aiming for a real-world deployment in collaboration with CMU Africa). In this project, we will build a set of unique applications that serve rural communities (e.g., online farmer markets, distance education). We will also investigate in what capacity Machine Learning-based applications can play a role in servicing such rural communities. # Doctor Recommendation Application for Rural Communities Nowadays, cloud computing is widely used to build web applications. However, while efficiency metrics (e.g., performance and energy) are critical to gauge cloud systems, we must start thinking about the societal implications of the web systems we build. This project represents the initial stage of developing and profiling a cloud application for underserved rural communities while considering intersectionality, equity, and fairness as first-order system design metrics. As a result, we can develop a framework for deciding the software and hardware designs under different constraints. For an initial effort, we plan to develop a cloud-based application to provide doctor recommendation services to rural African residents. It should be able to recommend doctors based on the health records of patients. In rural areas, however, there are many constraints, such as poor Internet connectivity and device capabilities. Therefore, we must carefully analyze the data storage and interaction models of each service, so that we may provide different services depending on the situation. In the case of poor networks, patients should be able to access a baseline recommendation service that returns all doctors. Under a decent network, advanced services, such as AI recommendations, should be available. In addition to designing a flexible application for rural communities, we will also create a local testing environment to simulate the rural situation. We will profile and evaluate our system's performance based on different assumptions about software and hardware. This application is aimed at being deployed at CMU-Africa, which will allow us to refine our system and application designs based on real-world workload and network patterns. ## What we will do - Discuss the system design and clearly identify the requirements of each service. - Build the application alone with a local testing environment. - Explore the software and har …