Welcome to delta-fusion’s documentation!

Delta (the a D aptive near r E a L - T ime A nalysis framework) facilitates near real-time streaming and analysis of big fusion data on remote HPC resources. It implements software components for loading and staging measurement data, streaming, analysis, and storage. Delta combines these components into a framework for adaptive near-real time analysis of streaming data, targeted towards use-cases in fusion energy research. It consists of multiple executables that send, receive, and process data on different machines. The picture below gives an overview of Delta.

Target architecture of Delta

At the data generation site, a generator reads measurement data from file and stages it for streaming. The data stream is received by a middleman that forwards it to the processor, which executes data analysis kernels on a supercomputer. The analysis results are stored and made available to web-clients by a separate webserver.

Running Delta

Describes how to install, configure, and launch distributed analysis workflows.

Running Delta

• Setting up to run Delta
  • Prerequisites
  • Obtaining Delta
  • Running pymongo in an MPI environment
• Configuring data analysis workflows
• Launching data analysis workflows using Delta
  • 1-node scenario
  • 2-node scenario
  • 3-node scenario

Package reference

Documentation of individual software modules

Package reference

• Data sources
• Data models
• Streaming
• Pre-processing
• Data analysis
• Storage

Presentations and Papers

Delta has been presented at multiple scientific conferences.

SciPy 2020

• Proceedings paper

• Presentation

23rd Conference on High Temperature Plasma Diagnostics

• Presentation

SC 2020 Workshop

2020 IEEE/ACM HPC for Urgent Decision Making Workshop

• Paper

Indices and tables

• Index

• Module Index

• Search Page