Autonomous Multi-Cloud Serverless...
summary
Melodic is an open source middleware that autonomously models, deploys and continuously reconfigures multi‑cloud applications across OpenStack/AWS/GCP/Azure, managing VMs, containers, serverless functions and FPGA nodes. Application requirements are written in Camel modeling (a TOSCA‑like DSL); optimization uses five solvers (Monte Carlo search, supervised+reinforcement learning, parallel tempering, learning automata) and a Morphemic forecasting extension for proactive adaptation. The demo deploys a Spark genome app with one spark master + 1–10 workers, starts at 1 worker (8 cores), then reconfigures to add capacity (16 cores); created VMs appear in AWS and monitoring shows the predicted load and triggers.
AI-written summary · about
chapters (11)
- 0:00Introduction to Melodic
- 13:02Live demo introduction
- 14:31Platform configuration steps
- 21:13Deployment process overview
- 24:24Constraint problem generation
- 28:56Reasoning and optimal solution
- 32:36Initial deployment completion
- 35:51Verification in AWS console
- 36:38Reconfiguration process
- 37:24Monitoring and metrics analysis
- 42:09Q&A
related talks
27mFrom Supercomputing to Serving: A Case Study Delivering Cloud Native Foundation Models
36mMLOps from the edge to on demand visualization
35mAn AI/ML-Driven Approach to Platform Lifecycle Management
21mHarnessing Microservices for Scalable, Fault-Tolerant ML Systems
27mGen-AI at Scale: Simplifying Orchestration of Healthcare Applications Across Multi-Cluster Environme
36mCatch Me If You Can – Leveraging Cloud Foundry, SDN, BOSH and Predictive Analytics
35mCloud-Native MLOps Platform for Python- and ML Engineers
48mBuilding Serverless AI powered Applications on AWS