CACHE-on

Mahindra École Centrale | India

Abhinav Reddy Nimma | Rohith Gilla

CACHE-On is a platform designed to solve the problem of internet penetration in rural areas. The idea, in a nutshell, is to utilise the in-place transport infrastructure to transfer user data as opposed to other delivery mechanisms (UAVs, Helium balloons) being explored by tech giants like Google and Facebook. The advantage of the approach proposed here is the minimal investments required given that the delivery network already exists in form of physical transport. We also develop two important applications based on this platform Edu-CACHE-On and Medi-CACHEOn. We chose these applications as a way of demonstrating the capabilities of the CACHE-On platform. For the former, the data flow is mainly from towns with Internet connections to the villages that lack Internet access, whereas for the latter, the direction is reversed. The platform and the applications built on them will use Machine Learning solutions at different levels to minimise redundancy in caching of data and to maximize its pertinence. The platform CACHE-On uses ML for intelligent caching and indexing of the cached content. The application Edu-CACHE-On uses Natural Language Processing (NLP) techniques to understand user queries as well as summarising the information gathered over the internet. For Medi-CACHE-on, we propose to use ML techniques to cluster medical data given by patients and highlight the cases with high probability of diagnosis to a Doctor found in the city centre. This will increase the efficiency of the doctor, as well as help quickly identify epidemics in remote places with no access to doctors.Bringing internet connectivity brings significant advantages to the villages and this newfound access can be leveraged amongst advertisers or service providing applications. Edu-CACHE-On and Medi-CACHE-On are examples of model applications, but the potential applications are plenty. We build an open platform which provides developers and applications with the new clients and their demands in mind. The more the number of applications that succeed, the stronger the platform will become due to the virtuous network effects. This will help in creating a self-sustaining business model, while also subsidising the end user.