Octopus
Our research paper, “Octopus: Experiences with a Hybrid Event-Driven Architecture for Distributed Scientific Computing”, has been accepted for presentation at the International Workshop on Fault Tolerance for HPC at eXtreme Scale (FTXS). Scientific research increasingly depends on distributed resources from HPC and cloud systems to edge devices. Event-driven architecture (EDA) enhances applications targeting these infrastructures by managing communication, processing, and security of distributed events. Enter Octopus—a hybrid cloud-to-edge event fabric designed to scale with demand, enable resilient applications, and enforce fine-grained access control. Supporting up to 9.6M events/second, Octopus is tailored for use cases like self-driving labs, data automation, and dynamic workflows.