Design foundations: why this product works the way it does

Most network monitoring tools work by watching for things they have been told to watch for. An engineer configures a threshold — say, CPU utilisation above 80% — and the tool raises an alert when that threshold is crossed. This approach has two structural problems.

The first is noise. In a real data centre, hundreds of thresholds are crossed every day for entirely benign reasons: scheduled backups, maintenance windows, burst traffic from a batch job. Every one of those events generates an alert, and the team learns to ignore them. When something genuinely significant happens, it arrives in a sea of noise.

The second problem is context. A threshold alert tells you that something happened. It does not tell you where in the network the root cause sits, what else might be affected, or how the event relates to traffic that was flowing at the time. Think about how your team investigates an alert today — how many tools does that typically involve, and how long does it take to build the full picture?

DC Assurance addresses both problems. It uses the structural knowledge inside Apstra Data Center Director — the full relational model of the fabric — to give every event context from the moment it is detected. And it learns what normal looks like in your environment, so it can distinguish genuine anomalies from expected variation without generating constant noise.

Any analysis is only as good as the data it works from. DC Assurance works from Apstra Data Center Director’s graph database — a structured, relational model that knows not just that a device exists, but what it is connected to, what it is supposed to be doing, and how its configuration relates to every other element in the fabric.

A traditional monitoring tool sees a stream of metrics. DC Assurance sees a network.

This is why the two approaches produce such different results. When your team gets an alert today, how long does it typically take to work out whether it is real and where it is actually coming from? For most teams, the answer is measured in hours. DC Assurance changes that to minutes, because the context that normally takes hours to assemble is already there.

DC Assurance runs on the same cloud platform as HPE Mist — a platform with more than ten years of operational experience running AIOps across wireless and wired enterprise networks at global scale. The infrastructure it relies on is running in production at some of the largest enterprises in the world.

AIOps for data centre networking is not experimental. The platform is mature, the data ingestion and machine learning pipelines are battle-tested, and the assurance patterns DC Assurance uses for data centre fabrics draw on the same foundational approach that has been refining itself for years in the wireless space.

AIOps — artificial intelligence for IT operations — is a term that gets used broadly and sometimes loosely. In this context, it means something specific: DC Assurance uses machine learning to establish a baseline of normal behaviour for your fabric, then continuously monitors against that baseline to surface deviations that a rules-based tool would miss.

This guide covers the AIOps capabilities available in the product today — root cause analysis and Service Level Experience scoring — and the reference section addresses common questions about what is available now versus what is on the roadmap.