Life beyond distributed transactions on the edge

Abstract

Edge/Fog Computing is an extension to the Cloud Computing model, primarily proposed to pull some of the load on cloud data center towards the edge of the network, i.e., closer to the clients. Despite being a promising model, the foundations to adopt and fully exploit the edge model are yet to be clear, and thus new ideas are continuously advocated. In his paper on \Life beyond Distributed Transactions: An Apostate's Opinion", Pat Helland proposed his vision to build\almost innite" scale future applications, demonstrating why Distributed Transactions are not very practical under scale. His approach models the applications data state as independent \entities" with separate serialization scopes, thus allowing ecient local transactions within an entity, but precluding transactions involving dierent entities. Accessing remote data (which is assumed rare) can be done through separate channels in a more message-oriented manner. In this paper, we recall Helland's vision in the aforementioned paper, explaining how his model ts the Edge Computing Model either regarding scalability, applications, or assumptions, and discussing the potential challenges leveraged .