PTero: Architecture and Development
summary
Describes a working workflow system PTero that encodes DAGs and runtime choices as petri nets, claims an amortized constant‑time step decision algorithm, and runs as horizontally scalable RESTful services (JSON, Redis, RabbitMQ, PostgreSQL). Shows concrete job backends (shell and LSF), a Vagrant development VM, CI and production monitoring, and reports deployment since 2013 that orchestrates 600,000+ workflows and 8 million+ tasks. Watch if you want an implemented, measurable system and integration details.
AI-written summary · about
chapters (10)
- 0:00Introduction to Taro
- 2:16Genomic data challenges
- 5:53Genome modeling system
- 7:56Legacy workflow engine
- 10:43Petri net architecture
- 15:45From flow to Taro
- 18:25Taro architecture details
- 26:56Practical workflow example
- 34:51Development and deployment
- 39:34Conclusion and Q&A
related talks
50mFault tolerant workflow orchestration on PHP
35mOrchestrating success
47mData Mesh in Kubernetes
51mMicroProcess Architectures: Planning for Change, & Retaining Flexibility
27mArchitecture Evolution of a Startup within an Enterprise
01h 00mFrom Planning To Production-Ready RAG
01h 00mReliable Agentic Systems need Durable Execution 🤖
22mManaging Distributed Workflows at Scale