Controls, Monitoring, and Operational Telemetry Guide

Introduction

A well-designed controls and monitoring system transforms a data center from a collection of independent equipment into a coordinated, observable, and manageable facility. This guide covers the planning and design of sensor networks, alarm systems, building management system (BMS) integration, data center infrastructure management (DCIM) platforms, and the operational telemetry architecture that ties them together.

Sensor Planning Methodology

Sensor placement should be driven by operational requirements, not equipment convenience. The methodology starts with identifying what operational decisions each sensor reading will support, then works backward to determine the required measurement points, accuracy, and update frequency.

Environmental Sensors

• Temperature: Rack inlet/outlet (every 3rd rack minimum), supply/return air at air handlers, outdoor ambient, and hot/cold aisle containment points. Accuracy requirement: +/- 0.5 degrees C.
• Humidity: Room-level relative humidity at supply air delivery points and critical zones. Accuracy requirement: +/- 3% RH.
• Airflow/Pressure: Differential pressure across containment boundaries, under-floor or plenum pressure, and fan discharge velocity at air handlers.

Electrical Sensors

• Power Metering: Main utility service, each transformer secondary, UPS input/output, distribution panel mains, and rack PDUs. Accuracy per IEC 61557-12 Class 0.5 or better for billing-grade measurements.
• Power Quality: Voltage, current, frequency, power factor, and harmonic distortion at critical points in the distribution hierarchy.

Leak Detection

Rope-type leak detection sensors along all water piping routes, under CRAH units, around chilled water connections, and in any area where fluid leakage could affect IT equipment. Zone-type controllers to localize leak position to within one meter.

Security and Access

Door contact sensors on all controlled entry points, card reader transaction logging, CCTV camera integration for visual verification, and cabinet-level access monitoring for sensitive racks.

Alarm Hierarchy Design

An effective alarm system requires a structured hierarchy that prevents alarm fatigue while ensuring critical conditions receive immediate attention. GridCore recommends a four-level alarm hierarchy:

• Level 1 - Critical: Conditions requiring immediate operator action to prevent equipment damage, data loss, or safety hazard. Examples: UPS on battery, fire detection activation, cooling system total failure. Notification: audible alarm, SMS, phone call, dashboard alert.
• Level 2 - Major: Conditions indicating significant degradation that will become critical without intervention. Examples: redundancy loss, approaching thermal limits, generator fail-to-start. Notification: SMS, email, dashboard alert.
• Level 3 - Minor: Conditions requiring attention during normal working hours but not immediate response. Examples: sensor drift, maintenance due, capacity thresholds. Notification: email, dashboard alert.
• Level 4 - Informational: Status changes and operational data logged for trending and analysis. Examples: scheduled maintenance events, capacity utilization milestones, environmental trending. Notification: dashboard log only.

BMS Integration

The Building Management System provides centralized monitoring and control of mechanical and electrical infrastructure. Integration design must define communication protocols (BACnet IP is recommended as primary, Modbus TCP as secondary), point naming conventions, polling intervals, and failover behavior.

Key integration points include: HVAC system control and monitoring, electrical switchgear status and metering, generator control and monitoring, fire alarm system annunciation (read-only), access control system event logging, and lighting control.

DCIM Platform Architecture

The DCIM platform provides IT-focused infrastructure management including capacity planning, asset management, change management, and operational analytics. It typically sits above the BMS layer and correlates facility data with IT workload information.

Architecture decisions include: on-premises vs. cloud-hosted deployment, integration with IT service management (ITSM) platforms, data retention and archival policies, reporting and dashboard customization, and role-based access control for different operational teams.

Operational Telemetry Architecture

The telemetry architecture defines how data flows from sensors through collection, processing, storage, and presentation layers.

• Collection Layer: Protocol gateways, data aggregation controllers, and edge processing nodes that normalize sensor data into a common format.
• Transport Layer: Network infrastructure (dedicated monitoring VLAN recommended) carrying telemetry data from collection points to central processing.
• Processing Layer: Real-time event processing engines that evaluate alarm rules, calculate derived values, and trigger automated responses.
• Storage Layer: Time-series databases optimized for high-frequency infrastructure data with configurable retention policies.
• Presentation Layer: Dashboards, reports, and mobile interfaces tailored to different operational roles (NOC operators, facility managers, capacity planners, executives).