Skip to content
Open access

High-Availability Architecture Patterns for Enterprise Network Management Platforms

Nov 2026 · International journal of advanced engineering science and information technology · 0 citations

Abstract

The activities of telemetry-based network operations require continuous collection of increasingly precise measurements. Making operations effective requires creation of a complete and trackable route between telemetry creation and source semantics and analysis, decision-making, remediation, and verification of results. The architecture described in this paper serves this need by integrating a source-aware collection, transport reliability, canonical normalisation, multi-tier processing, evidence-based findings, and remediation that is governed by policies. Topology-aware correlation and event-time processing can provide the context by which it is possible to distinguish between real and delayed, duplicated, or unrelated observations. Well-defined data-quality indicators: Clear data-quality indicators and metrics are necessary to lessen the confidence of analytics with missing or incomplete telemetry rather than being based on the interpretation of an indication of an operational state. Further separation of fast, operational, and planning analytics enables various control loops to run based on their latency, computation, and risk sensitivity. Adaptive telemetry is another approach for balancing observability and resource efficiency. In situations where additional evidence is justified, a wide coverage is achieved by baseline collection, and the diagnostic and forensic modes are selective to narrow the data to a finer degree. Adaptive collection lacks the possibility of fairness limitations and guided transitions that would provide false feedback or take control over available resources. This is further divided between automated action and the results of analysis regarding the architecture. Evidence-based discoveries sum up paralleled observing and policy regulation verification to approve, trust, fulfils a deadline, blast radius, rollback claim, and file validation requirement before remediation. The approach complies with automation and warns against operating responsibility. In general, the idea of telemetry is most beneficial when gathering is geared toward clear operational choices as opposed to the dashboard or unlimited building of data. Future areas of work can delve deeper into the architecture by learning-based anomaly detection, more advanced models of confidence, auto-optimization of the policy, and standard evaluation of remediation results in a wide variety of network environments.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.