The In-Memory Heart
Netcool/OMNIbus tracks alert information in a high-performance, in-memory database — the ObjectServer — and presents it through filters and views built for the operations floor. Events land in the ObjectServer's alerts.status table, where they are managed, correlated, and displayed in real time.
The moment a repeated event arrives, the ObjectServer collapses it into the single alert already on record, updating an occurrence count and timestamps in place rather than adding another row. That deduplication is the first and most important act of noise reduction — before a single correlation rule ever runs.
The Collection Fabric
A probe connects to an event source, acquires its data, and forwards it as an alert. Its rules file parses raw data into fields and builds an identifier that uniquely fingerprints the problem source. The library spans SNMP traps and informs, syslog, the Event Integration Facility, message bus and Kafka, and a broad range of vendor- and technology-specific integrations.
The in-memory database server at the platform's core. Probes and gateways forward events to it; it stores, deduplicates, and serves them to the event lists and every downstream layer.
Gateways replicate and route events: ObjectServer-to-ObjectServer for high availability, JDBC to external databases, a bidirectional gateway to ServiceNow for ticketing, and a message-bus gateway feeding analytics.
Repeated events from the same source become a single alert whose recurrence count and timestamps update in place — collapsing an event storm into one line an operator can actually act on.
Built to Stay Up
Large deployments run a multi-tier architecture — collection, aggregation, and display — each a primary/backup ObjectServer pair kept in lockstep by gateways, with automatic failover and failback.
In-memory ObjectServer for real-time event processing
Collection, aggregation, and display tiers for scale
Primary/backup ObjectServer pairs at every tier
Bidirectional gateways keep pairs synchronized
Automatic failover and failback — no single point of failure
AccuOSS designs and implements the probe-and-gateway fabric, sizes the tiers, and proves the failover — so the event core holds under the worst night of the year.