Home

BI|2020/107 Stateful Functions is an open source framework that provides a powerful abstraction for event-driven architectures, reducing the complexity of building and orchestrating distributed stateful applications at scale, by bringing together benefits of stream processing and Function-as-a-Service (FaaS). This work explores leveraging Stateful Functions in Flink in order to design and develop an extension to this Big-data platform to achieves increased scalability and performance, with adaptable resource management driven by service-level objectives regarding performance (latency, throughput), result accuracy, and resource management.

BI|2020/108 Graphs processing systems provide a powerful abstraction for sophisticated Big-Data applications. This work addresses graph storage (e.g., DGraph, Neo4j, JanusGraph, Nebula Graph) and processing (e.g., GraphX on Spark and Gelly on Flink) and their space- and time-efficiency, exploring how these metrics can be improved, e.g., by employing summarized data representations to reduce storage occupation, while allowing dynamism (i.e., update-able data), while still enabling implementing popular graph algorithms and workloads.