Workflows Information Flows Prof. Silvia Olabarriaga Dr. Gabriele Pierantoni
Content multi-layered structures and information flows (slide 1 in other pptx, motivation from e-science paper 2014) concrete example to illustrate: helio? mention that this is the same neuroscience gateways, PM separating and combining information flows: layers and APIs (slide 2 in other pptx with concrete example) challenges and possible directions ahead (aspect programming and semantic web, maybe mention PM)
Introduction & Motivation Integration among DCIs, workflow engines, portals and communities defined software Different user profiles: administrators developers users Increasing complexity in managing the information flows that connect components and users Work started during ER-FLOW and SCI-BUS A related paper presented in eScience 2014
Introduction & Motivation Integration among DCIs, workflow engines, portals and communities defined software Different user profiles: DCIs: administrators Communities defined software: developers and users Workflows: developers and users Workflow engines: developer, administrators and users Generic Portals: developers, administrators and users Portlets and custom build portals : developers, administrators and users Increasing complexity in managing the information flows that connect components and users Work started during ER-FLOW and SCI-BUS A related paper presented in eScience 2014
Overview Workflow Infrastructure Gateway Domain Experts Gateway Experts Workflow Experts DCI Installation, configuration, monitoring, troubleshooting Customized graphical User Interfaces Graphical User Interfaces, scripts, workflow design and debugging Workflow design and debugging API, development libraries and frameworks, code debugging
Overview
Layers and Users LayersUsers Community Gateways developers, administrators and users Generic Gatewaysdevelopers, administrators and users Workflow enginesdeveloper, administrators and users Workflowsdevelopers and users Communities Software developers and users DCIsadministrators
Layers and Information Flows DCI and External Services (DSI) Workflow Infrastructure Portal Infrastructure API/Commands GUI API/Commands GUI Actions Inform ation Community Gateway
The Heliophysics experience DCI and External Services (DSI) API/Commands GUI Specialized Standalone GUIs for Specific services Grids (gLite) & Clusters Workflow Infrastructure API/Commands GUI Specialized Web Services Interfaces Workflows and Meta-workflows TAVERNA GUI to execute jobs Portal Infrastructure API/Commands GUI WS-PGRADE Easy Interface to rapidly prototype GUIs HELIOGate Specialized Interface
The Heliophysic s example DCI and External Services (DSI) Workflow Infrastructure Portal Infrastructure Job Execution Results & Error Codes Workflows Execution Results & Error Codes HELIOGate User Defined Execution Results & Error Codes Workflow Development & Execution Prototyping and Development User Defined Execution
Workflows WS-Pgrade Metaworkflow TAVERNA metaworkflow TAVERNA Worfklows Execution in WS- Pgrade enabled portals Coarse Interoperability through the SHIWA Platform Execution on the TAVERNA line command tool
Solutions and future directions Present Solutions: there are different technologies and we cannot re- invent the wheel, so we must use the tools at disposal: Automatic test procedures triggered before execution to exclude some errors Automatic test procedures triggered on errors to determine if the infrastructure run well (does not solve all problems) Future Solutions, trying to develop systems at each layer where the various informations flows are tagged and dispatched automatically to the right component/user Information is expensive: we cannot keep all of it forever.