Skip to content

Cortex

Cortex is CobaltCore's intelligent placement and scheduling engine. It collects real-time state from across the infrastructure, maintains a knowledge database of cluster conditions, and uses that data to make smarter placement decisions for compute, storage, and network workloads.

What it solves

Nova's default scheduler makes placement decisions based on static filters and weights. Cortex enriches those decisions with live data - current hypervisor load, VM distribution patterns, storage latency, network topology - to optimize for resource utilization and workload performance.

Architecture

Components

ComponentRole
Retrieval & EnrichmentCollects and processes data from Kubernetes, Nova, Neutron, Cinder, and other sources
Knowledge DatabaseCentral store of enriched infrastructure state - the basis for all placement decisions
Initial PlacementHandles placement requests for new workloads; returns a placement suggestion to the caller
SchedulerContinuously monitors running workloads and triggers migrations when conditions change
APIExposes the knowledge database and placement engine to Aurora and operators

Key features

  • Modular plugins - data sources and scheduling algorithms are pluggable; Cortex adapts to different environments
  • Cross-domain - handles compute, storage, and network placement independently or as coordinated decisions
  • Production-scale - designed for thousands of placement requests per second in large cloud deployments

See also

EU and German government funding logos

Funded by the European Union – NextGenerationEU.

The views and opinions expressed are solely those of the author(s) and do not necessarily reflect the views of the European Union or the European Commission. Neither the European Union nor the European Commission can be held responsible for them.