作者: Emanuele Carlini , Patrizio Dazzi , Alessandro Lulli , Laura Ricci
关键词:
摘要: In this day and age, human being is generating an unprecedented amount of data coming from multiple heterogeneous sources, ranging scientific devices to business transactions. many contexts such modelled as a graph, however, due its size, it often infeasible process on single machine. As consequence, the adoption parallel distributed computing frameworks natural approach face problem. This paper proposes high-level for large graphs processing. Our takes inspiration overlay networks, widely exploited information dissemination, aggregation orchestration in highly systems. We propose Telos, programming framework supporting definition computations composition overlays, each devoted specific aim. Telos implemented top Apache Spark. A set experimental results are presented give preliminary evidence effectiveness our approach.