From Events to Continuous Intelligence – The New Frontiers of Unbounded Stream Computing
We warmly welcome Paris Carbone. He is a researcher in computer systems, currently leading the efforts of the continuous deep analytics group at RISE and KTH, an effort to integrate data computing workloads under a unified architecture for programming, compilation and execution.
Date and time: November 26 2020, 12pm – 1 pm
Speaker: Paris Carbone
Title: From Events to Continuous Intelligence – The New Frontiers of Unbounded Stream Computing
Zoom: https://kth-se.zoom.us/j/67432682790?pwd=dVgzbjRSbUVFT2FOYTByYlZrTU9BUT09
Meeting ID: 674 3268 2790
Password: DF2020
Abstract: Stream processing has been an active research field for decades and is currently put to use today at extreme scale to support real-time stateful services and analytics. In this talk, we will first cover some of the most major findings and steps that allowed for the evolution of streaming systems. Stream processors have matured from experimental prototypes to fully reliable and automated systems while subsuming application use cases previously addressed by specialised architectures including large-scale continuous ETL and analytics, complex event processing, ML model serving and event-driven cloud applications.
However, are stream processors equipped for the next wave of AI and IoT applications? In the second part of this talk we are going to explore a set of new frontiers in unbounded computing imposed by new types of applications such as graph and tensor streams including code generation for modern hardware, workload-aware state management, new programming models and support for decentralised pipelines among others.
Bio: Paris Carbone is a researcher in computer systems, currently leading the efforts of the continuous deep analytics group at RISE and KTH, an effort to integrate data computing workloads under a unified architecture for programming, compilation and execution. Paris holds a PhD in distributed computing from KTH and is one of the core committers for Apache Flink®, the most popular stream dataflow system to date with key contributions to its state management.