Monitoring
A BMS (building management system) runs a building: its heating, cooling, lighting, security and life-safety systems. DCIM (data centre infrastructure management) manages the physical assets and resources of the data centre itself: power down to the rack, space, capacity and the equipment inventory. Both can read the same UPS, and the real question is which one owns its alarms.
A BMS controls the building; DCIM manages the data centre's physical infrastructure and its capacity. Lawrence Berkeley National Laboratory describes DCIM as a category of solutions created by the IT industry to extend data centre management to all physical assets and resources in the facilities and IT domains, and a BMS as a system that works toward managing the entire building, including lighting, cooling, power and security.
The practical difference is direction. A BMS is built to control: it starts pumps, changes set points and runs sequences on the mechanical plant. DCIM is built to see and plan: it holds where each asset is, what it draws, how much capacity is left and what a change would do. In most data centres both exist, and both want data from the same UPS, PDUs and cooling units.
DCIM gathers data from many sources into one record of the physical data centre. LBNL's guide lists what that record holds: energy use, location, asset life cycle, environmental conditions, capacity, connectivity and configuration. DCIM tools present that data to decision makers, mostly graphically, and some can model proposed changes so that facilities and IT teams see the effect on capacity before they make them.
For the power path, that means DCIM knows which UPS and which PDU feed a rack, how loaded each one is, and whether a new server fits within the redundancy the room was designed for. It is also the natural home for PUE tracking; see PUE explained.
A BMS supervises and controls building systems: heating, ventilation and air conditioning, lighting, security and life safety. In a data centre it usually runs the chillers, pumps and air handlers, and it watches the electrical plant at a high level. LBNL's guide gives integration with the building automation system for fault detection and diagnosis as an example of what joining energy data to the BMS makes possible.
Where a BMS is weaker is inside the white space. It sees the cooling plant and the main electrical gear well, but not usually the rack, the server or the individual battery block.
Through the UPS network card, using different protocols for each system. LBNL's guide asks that monitoring cover power (UPS, PDU), cooling, generators, servers and sensors over several protocols: BACnet, Modbus and SNMP. As a manufacturer example, Vertiv's Liebert IntelliSlot Unity card monitors a wide range of UPS operating parameters, alarms and notifications, and communicates with building management systems and network management systems over BACnet, Modbus and SNMP. Battery monitors work the same way: Vertiv's Albér monitors interface to third-party building management systems over Modbus, and offer BACnet and Modbus through the same card family.
Two settings in Vertiv's card guide matter whichever system connects. The card can be set to read-only for BACnet and Modbus, so a building system cannot change the UPS; and the BMS should be configured to send no more than two writes per second on a sustained average. The SNMP community strings should be changed from their defaults before the card goes on the network.
Both can display them, but the alarm itself should go straight to the people who respond, not wait in either console. A BMS is often watched from a building operations desk that may not know what a UPS on bypass or a battery string in alarm means. DCIM may be watched only in office hours. Schneider Electric's guidance applies here too: each alarm needs its recipients, its route (email, SMS, SNMP trap) and an escalation path when it is not resolved in time.
A workable split is common: the BMS keeps the cooling plant and building alarms; DCIM keeps capacity, asset and PUE data; and the UPS and battery monitor send their critical alarms directly to a monitoring service as well. See remote UPS monitoring and alarm response for how that last path is set up.
| Site | Usually enough | Why |
|---|---|---|
| One server room with a single UPS | UPS network card, a few room sensors, alert routing | Few points; the alarms matter more than dashboards |
| Building with its own mechanical plant | BMS for the plant, plus direct UPS and battery alarms | The BMS already runs cooling; the UPS needs a responder |
| Multi-room or multi-site data centre | DCIM for assets and capacity, BMS for the plant | Capacity and change planning across many racks and feeds |
Whatever the size, the UPS and battery alarms come first. A site with no DCIM and a well-routed UPS alarm is better protected than one with a full DCIM platform that nobody is watching overnight. Monitoring integration, including BACnet or Modbus mapping into an existing BMS, is delivered through specialist subcontractors and quoted with the UPS service; UPS maintenance and repair services cover the equipment side.
No. A BMS controls building systems such as HVAC, lighting and security. DCIM manages the data centre's physical assets, power, space and capacity. They often share data.
Yes, through the UPS network card over BACnet or Modbus. Set the card to read-only for those protocols so the BMS cannot change the UPS.
Mostly SNMP for IT and power equipment, with Modbus and BACnet for facility equipment. LBNL's guide calls for SNMP, Modbus and BACnet over TCP.
No. Critical UPS and battery alarms should also go directly to the people who respond, with escalation, because a building operations desk may not know how to act on them.
Sources, checked 2026-10-04:
Tell us the site, the UPS and the timing. We reply and quote within 4 business hours, or call 1 (438) 800-0382
Brand names are trademarks of their owners and are used only to identify the equipment we service. We are not the manufacturer and are not affiliated with it.