Serverless Edge Providers for AI Applications

In present times, cities and industries have the opportunity to improve with emerging paradigms such as Smart Cities and Industry 5.0. The use of new technologies like Artificial Intelligence and Cloud Computing brings new applications that have the chance to improve quality of life in general. This paper refers to serverless edge computing as a means to address the increasing data transfers and computations associated with the Internet of Things in Society 5.0. The benefits of this computing model are discussed, such as reduced latency and cheaper computations, and various architectures and technologies related to this paradigm are examined. The emergence of serverless edge providers is also explored, which shows a trend on integrating edge computing with existing infrastructure. To this end, an extensive review has been performed on different serverless edge providers, aiming to study how the service is offered. The methodology used consists in a review of their documentation, contrasted with existing academic literature. The findings show limitations on what providers currently offer, due to them being based mostly on preexisting Content Delivery Network infrastructure. A business model is proposed that focuses on providing serverless edge services within cities. The proposal consists in distributing low-power edge servers throughout the city and utilizing wireless networks for connectivity. The results would be lower latency, and easier application development. Serverless edge computing has great potential as an enabler of smart city applications, which highlights the need for providers operating within cities.