IVOA Grid & Web Services: Remote Execution

Contents

Overview

The size of astronomy data, already significant, will grow rapidly with the upcoming large projects. The tools available for astronomers will be crucial for their research. With technologies like Docker components and notebooks like Jupyter, it will be possible to build highly flexible working environments (both on the user and on the server sides) and to customize them quickly as needed. One of the fundamental stakes is to allow astronomers to carry out their analyses on these masses of data without having to transfer them to their own location. For example, a Jupyter Notebook allows you to enter source code directly into a browser, to run it remotely, and to display images, graphics, etc... in the browser. This interactivity allows the astronomer to test and adapt quickly an algorithm with several ranges of parameters. For heavier processing, we can imagine the capability of sending code via components such as Docker to a data centre able to integrate it into one of its services, allowing users to experiment with their own processes where the data is located.

The Grid and Web Services is in the process of defining a standard means of remotely executing code.

• Docker and Containers
• API may describe:
  • Code
  • Parameters
  • Reference Data
  • Execution environment
• Code in Publications, reproducability of computing science results
• Process cannot not impede or dictate workflow
• Machine learning / big data / remove latency