GTC vs. DCIM – Data Center Monitoring

Data centers are unique in that they belong to both the construction industry AND the IT sector.

This dual nature has significant implications for maintaining operational readiness, particularly regarding real-time monitoring of the data center, with potential failure scenarios such as those described below:

• Removal of responsibility from the facilities and/or IT teams regarding data center monitoring
• Failure to communicate expected service levels, resulting in improper configuration of alerts or monitoring metrics
• Outdated monitoring tools that no longer reflect the reality of the data center, which has evolved over time
• The concept of redundancy is not taken into account in monitoring
• No predictive or capacity-based management of the data center

The mistake often stems from the belief that a single tool can meet the needs of all organizations. Based on our experience from over 120 data center implementations, in most cases, two separate tools are needed to ensure that the data center is properly monitored.

The first tool, designed for a company’s general services or technical departments, is based on a building management system (BMS). As an industrial-grade tool, it relies on proven, real-time protocols such as dry contacts, MODBUS, and BACnet.

The data reported consists mainly of Boolean alerts with some numerical information. We will implement advanced, static technical overviews (electrical diagrams, hydraulic diagrams, status of UPS components—including the inverter, batteries, and air conditioning units—as well as the status of valves, generators, etc.).

Simplified dashboards are available to external monitoring teams or internal operations centers. For this data center monitoring tool, we aim for speed, simplicity, and comprehensive reporting.

The second tool designed for CIOs is a DCIM (Data Center Infrastructure Management) monitoring tool, similar to those used for network monitoring. Relying primarily on SNMP messages, it provides a wealth of information on individual components (such as air conditioning systems, UPSs, etc.).

Instead, we will focus on monitoring trends, curves, and averages. The metrics are more complex (such as averages or temperature changes, PUE, etc.).

DCIM alerts can be more predictive because they are based on significant increases or pre-alert thresholds. Notifications can be sent to different people, with options to track the status of alerts and escalate them.

The business modeling of the data center performed in the DCIM (layout, density, and placement of cooling units) will also make it possible to analyze the constraints and interactions between the various technical systems and to plan for the future.

For this data center DCIM monitoring tool, we are looking for modeling capabilities, the ability to store large amounts of data over time, and advanced, flexible reporting features.