Where the spreadsheet breaks
Keeping a rack inventory in a spreadsheet works for a small server room. As the infrastructure grows, the same method quietly breaks — and you notice not when it breaks, but at the worst possible moment: a U that looks free in the list turns out to be occupied, an IP address gets assigned twice and takes down a production service, a switch port is disabled without knowing what depends on it, or one more device pushes the rack past its PDU capacity.
Data Center Infrastructure Management (DCIM) gathers all of this around a single source of truth.
DCIM is not a monitoring tool
- A monitoring tool (Zabbix, Nagios, Prometheus) answers is it up, how loaded is it? It collects time-series metrics.
- DCIM answers where is it, who owns it, what is it connected to, how much space and power does it use? It holds inventory and relationships.
The right setup combines both: DCIM holds the inventory and pulls status from the monitoring tool into one screen.
Core DCIM modules
Rack diagram
A visual 1U–48U map of each rack is the most used screen in DCIM. Devices are placed by drag and drop, each recording U height, make and model, serial number, tenant and power adapter type. Front and rear views are kept separately — cabling lives at the rear, and that distinction matters on site. Practically, it answers "which rack fits this 3U server?" without walking into the room.
Physical host → hypervisor → VM hierarchy
Physical inventory alone is not enough. DCIM should hold all three layers, and this data should be discovered rather than typed: vCPU count, RAM, datastore-level disk usage, network interfaces and IP addresses are pulled through the VMware vSphere API, Proxmox API and Microsoft Hyper-V WMI. That makes "which services are affected if this host goes into maintenance?" an instant question to answer.
IPAM — IP address management
IPAM solves the most urgent problems. IPv4 and IPv6 subnets, supernet/subnet hierarchy and assigned/reserved/free address views live in one place. The critical feature is DHCP synchronisation: lease data is pulled from ISC DHCP, Windows DHCP or MikroTik DHCP, and a warning is raised when a statically assigned IP collides with a lease — catching an outage before it happens.
Cable and port tracking
Physical links between patch panels, switch ports and device NICs are documented with cable type (Cat6, Cat6A, Fiber SC/LC), length and colour code. The real value shows up in impact analysis: disabling a port shows every affected connection graphically.
Power and cooling capacity
Smart PDUs (APC, Raritan, Vertiv) are read over SNMP for real-time consumption in W, kW and kWh, measured per rack, with alerts as maximum capacity approaches. From this, PUE (Power Usage Effectiveness) can be calculated:
PUE = total facility energy / IT equipment energy
1.0 is theoretical perfection; 2.0 means that for every 1 kW of IT load another 1 kW goes to cooling, lighting and losses. Tracking PUE over time is the most direct way to measure the return on cooling investments.
Change management
Adding, removing or moving hardware, changing an IP and updating configuration are bound to an approval workflow. Requester, approver, implementer and timestamp are written to an immutable audit log, and a rollback plan is documented. "Who approved this change?" must have a recorded answer.
Multi-site management
Organisations with several data centres keep an independent floor plan per location. An SVG-based editor draws the rack grid, marks hot and cold aisles and visualises inter-site links, so every location is visible from one screen.
How to start a DCIM project
- Start with physical inventory. It takes effort, but only once.
- Turn on discovery. Hand-entered data goes stale; discovered data does not.
- Bring IPAM online. Import existing IP tables and set up DHCP synchronisation.
- Make the change workflow mandatory. Changes made outside DCIM make DCIM untrustworthy.
- Add power measurement. PDU integration turns capacity planning from guesswork into measurement.
DCMon
DCMon, developed in house at Marta Teknoloji, delivers all of the above in one platform: 1U–48U visual rack diagrams, SVG-based multi-site floor plans, the physical host–hypervisor–VM hierarchy with automatic discovery for VMware vSphere, Proxmox and Hyper-V, IPv4/IPv6 IPAM with DHCP lease synchronisation, real-time SNMP v2c/v3 device monitoring, smart PDU integration with PUE calculation, approval-based change management and cable/port tracking. It integrates with ITSM tools such as ServiceNow and Jira, and with Zabbix and Nagios.
Get in touch to have your current data centre inventory mapped and to see DCMon in a demo environment.