Data center power, cooling & DCIM
Every server, switch and camera eventually depends on a UPS, a power distribution unit and a generator that starts when it is needed. The uninterruptible power supply (UPS) comes in three common designs. Standby (offline) units run from utility power and switch to the inverter when it fails, which can take up to about 25 ms. Line-interactive units add a transformer that corrects brownouts and surges without using the battery. Online double-conversion units rectify incoming power to DC and invert it again, so the load never sees the utility and there is no transfer switch.
What we do
- UPS topology and runtime sizing
- A and B feed and transfer-switch design
- Rack PDU metering and power-capacity planning
- Cooling and airflow review
- Generator and fuel test plans
- DCIM and UPS monitoring wired into alerting
- Power-event runbooks and shutdown orchestration
- Small server rooms, micro data centers and edge sites
- kW-per-rack budgets against feed capacity and cooling
- Inlet-temperature targets set from the ASHRAE TC 9.9 guidelines
We design for the failure, not the brochure: A and B feeds from separate sources to dual-corded equipment, rack PDUs with per-outlet metering, automatic transfer switches, generator and fuel plans, battery runtime that covers a clean shutdown, and cooling sized to the real heat load with hot and cold aisle containment. Capacity is planned in kilowatts per rack, not in square feet, and tested with load, not assumed.
Data center infrastructure management (DCIM) software tracks racks, power draw, temperature and assets so capacity decisions use data. Schneider Electric (APC and EcoStruxure IT), Vertiv and Eaton are the leading names in this field, and our team has worked with their equipment; we are not their partner or reseller. We connect UPS and PDU monitoring to your alerting, document runbooks for power events and test transfers on a schedule. Part of our network, server and managed IT practice.
Leading platforms we work with
- Schneider Electric
- Vertiv
- Eaton
- APC
- Legrand
- Raritan
- Server Technology
- CyberPower
- Tripp Lite
Platforms and practices
- Double-conversion, line-interactive and standby UPS
- Rack PDUs and metered outlets
- A and B power feeds
- Automatic transfer switches
- Generators and fuel planning
- Lithium-ion UPS batteries
- Hot and cold aisle containment
- Precision and in-row cooling
- Environmental sensors and leak detection
- Power and capacity monitoring
- DCIM software (EcoStruxure IT)
- Micro and edge data centers
- kW per rack budgeting
- ASHRAE TC 9.9 thermal guidelines
References
- Schneider Electric: data center solutions
- Vertiv: uninterruptible power supplies
- Vertiv: solutions
- Wikipedia: Uninterruptible power supply
- Open Compute Project: Modular Hardware System
- Wikipedia: Data center (A/B feeds, aisle containment, ASHRAE recommended inlet range)
Sources last checked .
Product names are trademarks of their owners and are listed to describe our experience; they do not imply endorsement or partnership.
A fabric with no single point of failure
Every leaf switch has a 100G QSFP28 uplink to every spine, servers attach at 25G or 10G through LACP pairs, and ECMP spreads flows across all spines. If a spine fails, the flows simply rehash onto the rest.
How we use Data center power, cooling & DCIM for clients
Frequently asked questions
Direct answers to the questions we hear most. Something else on your mind?
Send us a messageWhich UPS type do we need?
It depends on the load. Standby units switch to the inverter when utility power fails, which can take up to about 25 ms. Line-interactive units also correct brownouts and surges without using the battery. Online double-conversion units rectify and invert continuously, so the load never sees the utility and there is no transfer switch.
What does A and B power mean?
Two independent power feeds from separate sources, each supplying one power cord on dual-corded equipment. If one feed fails, the other carries the load. We design the feeds, rack PDUs and automatic transfer switches so that one failure, or one maintenance window, does not stop the room.
How much battery runtime do we need?
Enough for a clean shutdown at minimum, and enough to bridge to the generator if you have one. We size runtime from the real load, plan generator and fuel arrangements, and test transfers on a schedule. Untested batteries and generators are the usual cause of a power event turning into an outage.
Why plan capacity in kilowatts per rack?
Power and heat, not floor space, limit a modern room. We plan capacity per rack in kilowatts, size cooling to the real heat load with hot and cold aisle containment, and verify with load tests instead of nameplate numbers. Dense GPU servers make this much more important.
What does DCIM software do for a small site?
Data center infrastructure management (DCIM) software tracks racks, power draw, temperature and assets, so decisions about adding equipment use data. We connect UPS and PDU monitoring, such as EcoStruxure IT, to your alerting, and write runbooks for power events. Even a micro or edge site benefits.
Are you a Schneider Electric, Vertiv or Eaton partner?
No. Our team has worked with their equipment, and we advise on and operate what you buy. We are not their partner or reseller. Brand names on this page describe experience, and we select products by fit to your environment.