Oracle Unified Assurance · Architecture

Four layers, one source of truth.

Unified Assurance is built as clean, layered architecture — collection, processing, data, and presentation — wired together by cloud-native microservice pipelines that scale independently.

The Layers

Every layer does one job — and only one.

Collection layer

Inbound collectors and active pollers gather events, metrics, flows, and configuration from a multi-vendor estate — Syslog, SNMP trap and poll, webhooks, TL1, email, and file and TCP sources — and normalize them for everything downstream.

Cognitive Engine

The Unified Assurance Cognitive Engine is the processing tier — service-level management, the CAPE correlation engine, Root Cause Analysis, anomaly detection, thresholding, mechanizations, and ticketing, all operating on the live stream.

InfoLake data layer

Oracle Communications InfoLake is the single source of truth — purpose-built stores for real-time events, time-series metrics, the topology graph, historical analytics, and configuration state, each chosen and scaled for its job.

Presentation layer

A shared presentation tier delivers event lists, dashboards, diagrams, analytics, metric charts, network detail, and configuration diffs and history to the operations floor.

The Pipelines

Cloud-native microservice pipelines.

Work is organized into independent pipelines, each scaling to its own load.

Discovery

Discovers devices and registers them so every other pipeline can monitor and poll them.

Event

Collects and stores real-time events from discovered devices for fault management.

Metric

Collects performance metrics into the time-series store for monitoring and thresholding.

Topology

Builds and maintains the topology graph that powers root cause and Vision.

Flow

Enables Flow Analytics to monitor bandwidth consumption from network flow data.

Cloud-Native by Design

Stateless microservices, orchestrated to scale.

The collection and processing tiers run as stateless microservices in container-orchestrated clusters — scaling out under load, self-healing on failure, and spanning data centers for geo-redundancy.

  • Containerized, Kubernetes-orchestrated microservices

  • Independent, per-pipeline horizontal scaling

  • Event-driven autoscaling under load

  • Automatic failover and self-healing

  • Cross-data-center geo-redundancy

Architecture that scales with the network, not against it.

AccuOSS sizes and stands up the collection, processing, and data tiers for your estate — and proves the resilience before you rely on it.

Talk to our teamFault & Root Cause