RU Contact us
DC / 07 of 09 · Data centers

A data center where power, cooling and network are designed together

We set up and commission data centers. We work through the failure of every element before it happens.

When people call us
  • 01Redundancy exists on the diagram but has never been tested
  • 02Racks keep being added, power headroom is never recalculated
  • 03Servers in one row overheat while the cooling unit is fine
  • 04The documentation describes the data center of three years ago
+Data centers: ink illustration
PWR / both power feeds on one panel
HOT / hot air recirculation in row three
DC / 07Findings are illustrative
+
Day one

What we check first

  • Trace both power feeds from the rack to the utility entrance
  • Calculate load per rack, row and UPS under failure
  • Review airflow: blanking panels, aisle containment, recirculation
  • Compare the cable records and diagrams with what is on the floor
Tool

Power and heat of the server room

Current is for three-phase 400 V at a power factor of 0.95.

+
Failure patterns

Where it usually breaks

A

Two feeds, one point of failure

Both rack feeds end at the same panel or the same UPS. The redundancy lasts until the first failure.

B

Headroom eaten gradually

Each new rack adds “a little”. Two years on, one UPS runs beyond its limit when its neighbour fails.

C

Cooling that misses the servers

Empty units have no blanking panels. Cold air bypasses the servers and hot air returns to the intake.

+
Outcome

What you keep

  • A load and headroom model for power and cooling
  • A failure test report: power, cooling, network
  • Current as-built documentation
← MON

Monitoring

VND →

Cisco and other vendors

+
Contact

Describe your task

Tell us what is happening and what has been tried. The engineer who will handle it replies.