MiG - A Complete Grid Middleware (mostly) in Python
summary
MiG is a Python-first grid middleware that runs a centralized service layer while keeping no client install required on resources — resources create local user accounts and pull jobs over SSH inbound while users authenticate with x509 certificates via a browser. It implements per-user homes, runtime environments, a central software repo using zinstall packages for dynamic deployment, a web-based job/console UI, experimental VM integration, and small-scale production use for research projects rather than wide commercial deployment.
AI-written summary · about
chapters (16)
- 0:00MiG introduction and origins
- 2:12First-gen grid limitations
- 5:08MiG design principles
- 8:20User and resource management
- 12:33Job submission and runtime
- 16:01Virtual organizations and collaboration
- 20:03Resource models for clusters
- 23:01Why users moved to cloud
- 26:57Learning from cloud for MiG
- 31:10Building platforms in MiG
- 33:02MiG architecture and components
- 36:23Python and SSH integration
- 39:39Developing MiG in Python
- 41:58Current status and challenges
- 44:20MiG capabilities and use cases
- 46:47Q&A
related talks
31mInside ASGI The Engine Behind Modern Python Web Frameworks
25mA microservices architecture with Python and Docker containers
24mLightning Talks
57mDesigning a scalable and distributed application
39mArchitecture of a cloud hosting service using python technologies: django, ansible and celery
40mHomemade scientific clusters in Python
01h 34mPython for Distributed Systems
01h 49mBuilding Web-based Analysis & Simulation Platforms