Revamping Power, Cooling, and Distribution in Data Centres: A Journey from the Rack Upwards
Revamping Power, Cooling, and Distribution in Data Centres: A Journey from the Rack Upwards
TL;DR: Effective data centre infrastructure isn't just the combination of isolated components; it results from a series of decisions starting from the rack and moving outward. This guide helps IT leaders and data centre managers navigate the integrated selection process for racks, power distribution, UPS, and precision cooling, illustrating how each element influences the others and what to consider at each step.
Why is the Rack Where Every Data Centre Decision starts
Most infrastructure discussions begin incorrectly. Facility managers focus on UPS capacity, cooling systems, and power distribution without first understanding how the rack layer affects or limits these decisions. The rack is more than just passive furniture; it forms the structural and thermal basis that influences cable management, airflow, power efficiency, and system scalability. Choosing the right rack early on is crucial—it's not mere pedantry but the quickest way to prevent costly retrofits later during deployment.
In standard enterprise data centre settings, NetShelter SX enclosures offer essential features like structural strength, cable management, and airflow optimisation, forming a reliable base for subsequent infrastructure. These enclosures come in 42U sizes with depths ranging from 1,000mm to 1,200mm and are suitable for high-density setups. They seamlessly connect with aisle containment systems, transforming open-plan server rooms into controlled thermal environments. Hot aisle containment, which directs server exhaust into a dedicated return path instead of mixing back into the room, proves to be one of the most cost-effective efficiency measures without requiring changes to existing active infrastructure. Did You Know? Data centres that properly implement hot and cold aisle containment can usually cut cooling energy use by 20–40% compared to open-plan setups with the same IT load, all without any changes to the cooling equipment.
Power Distribution: From the Room to the Rack, and Why Intelligent Metering Transforms Everything
After establishing the rack layer, the next crucial decision is power distribution, which influences long-term efficiency and operational clarity. The debate over basic versus intelligent PDUs is often seen as a cost issue, but it is fundamentally a data issue.
Basic rack PDUs simply distribute power. In contrast, intelligent rack PDUs like the NetShelter Rack PDU Advanced go beyond that by measuring, reporting, and alerting. They allow real-time monitoring of each outlet's load, continuous tracking of rack-level energy consumption, and can use temperature and humidity sensors to trigger alerts before conditions become harmful to equipment.
For data centre managers overseeing multiple racks or remote locations, this visibility isn't a luxury; it's essential for transforming capacity planning from guesswork to data-driven decision-making. Overloaded circuits, underutilised capacity, and thermal anomalies—often unnoticed until they trigger downtime—are now detectable weeks in advance.
At higher-density deployments above 30–50 kW per rack, overhead busway systems replace traditional under-floor cable trays as the preferred power distribution architecture. Busway eliminates under-floor airflow disruption caused by cable penetrations, allows branch circuits to be added without downtime, and reduces the commissioning complexity associated with under-floor infrastructure changes. The practical implication: facilities that plan for busway during the initial build avoid the expensive, disruptive retrofits that under-floor cable tray systems eventually require as density increases.
UPS Selection: Aligning Protection Architecture with Workload Criticality
The UPS layer combines protection level, efficiency, and scalability into a unified infrastructure decision, and it is where the difference between an adequately specified and a correctly specified system has the most significant operational impact.
Three UPS configurations are used in data centre settings, each with its own advantages and disadvantages. Line-interactive UPS offers automatic voltage regulation and battery backup, with a transfer time of 2–4ms, suitable for less critical loads where a short downtime is acceptable during switching. Online double-conversion UPS ensures zero transfer time and provides full isolation from grid fluctuations, making it the preferred choice for racks housing servers, active storage, or network switching gear, where even a millisecond delay could cause data loss or service disruption.
The Galaxy VL and VXL series are the standard for efficiency and scalability in enterprise data centres. They achieve up to 99% efficiency in ECO mode during Live Swap module replacement—allowing power modules to be added or replaced without shutting down the system. These series eliminate the false trade-off between redundancy and efficiency that older UPS designs imposed on operators.
Modularity is especially important here. A UPS set to full capacity from the start involves substantial stranded capital expenditure in a facility that gradually reaches its design load. Modular UPS systems enable operators to start with the required capacity and expand by adding power modules as load increases, reducing initial costs while maintaining full scalability. For deployments anticipating significant workload growth over three to five years, this strategy generally offers superior capital efficiency compared to fixed-capacity options.
Precision Cooling: Aligning Thermal Density with Rack Density
Cooling depends on the decisions at the rack and power layers, which can either enhance efficiency or lead to operational issues. A well-designed rack with smartly monitored power distribution provides the cooling system with precise data and a stable thermal environment. Conversely, a poorly planned rack setup results in unpredictable hotspots that even advanced cooling systems cannot manage effectively.
For data centres with less than 20 kW per rack, precision InRow DX cooling in an N+1 setup offers dependable, energy-efficient thermal control. It captures heat at the source, unlike perimeter cooling units that rely on room airflow and are less effective at higher densities. This approach minimises energy use by reducing the need to move air across the data hall and promotes more consistent thermal conditions across the rack face. Above 20–30 kW per rack, which is becoming more common as server capacities grow, rear-door heat exchangers serve as an effective link between air-cooled systems and full liquid cooling solutions. For example, Motivair ChilledDoor units can cool up to 72 kW per rack by directly transferring exhaust heat to a chilled-water circuit at the rack door, without requiring modifications to the servers or extensive civil work. This makes them ideal for environments that gradually increase density rather than replacing infrastructure all at once.
The Integrated Stack: The importance of each Layer informing the Others
This journey begins with the rack, as each layer imposes constraints and offers opportunities for those above. A rack with poor cable management hinders airflow and undermines even the most advanced cooling systems. An unmetered power distribution setup prevents the UPS from having the necessary load visibility for optimal operation. Similarly, a cooling system designed for current rack densities without a clear upgrade plan may require a complete replacement rather than allowing phased expansion as workloads increase.
Choosing these four layers—rack, power distribution, UPS, and precision cooling—as a unified stack, rather than as separate procurement choices, distinguishes data centre infrastructure that operates predictably from that which constantly requires reactive adjustments. EcoStruxure IT connects this integrated stack at the software level, offering vendor-neutral monitoring of power and cooling systems, real-time visibility into racks, and predictive alerts that enable operations teams to preempt problems. The hardware choices laid out earlier create the physical base; EcoStruxure IT delivers the operational insights that ensure the foundation performs optimally. The journey from the rack up is literal, not metaphorical. It describes how infrastructure decisions influence each subsequent step. Making the right choices at each stage builds a data centre that delivers reliability, efficiency, and true scalability as new workload demands arise.
Frequently Asked Questions
The key power distribution solutions for modern data centres emphasise intelligent, metered visibility along with dependable power delivery, moving beyond simple connectivity. At the rack level, devices like the NetShelter Rack PDU Advanced offer outlet-specific load monitoring, live energy use data, and environmental sensors for temperature and humidity. At the room level, overhead busway systems allow for easy addition or reconfiguration of branch circuits without interruption, unlike under-floor cable trays that require significant disruption. At the facility level, modular switchgear and power monitoring tools such as EcoStruxure Power Monitoring Expert connect the entire electrical network with real-time health insights. Together, intelligent rack PDUs, overhead busways, and centralised power monitoring set the standard for data centre power distribution in mid-to-large enterprise setups.
InRow cooling involves precision units placed within server rows to capture heat directly from IT equipment, preventing it from mixing with room air. In contrast, traditional perimeter cooling uses CRAC or CRAH units at the room boundaries, requiring large air volumes to cool servers, which can create thermal inconsistencies and increase energy consumption. For rack densities below 5–8 kW, perimeter cooling with hot- and cold-aisle containment is effective.
However, for higher densities (15–20 kW and above), InRow cooling offers improved efficiency and thermal consistency. A hybrid approach, using perimeter cooling for low-density areas and InRow cooling for high-density rows, balances cost and performance in mixed-density environments.
Four indicators show when upgrading rack infrastructure offers better value than ongoing maintenance. First, if existing racks can't handle current cable volumes without affecting airflow—evidenced by hot spots, higher inlet temperatures, or uneven cooling—the cable management system is no longer suitable. Second, if racks are incompatible with current aisle containment systems, the facility misses out on 20–40% of cooling energy savings that containment provides. Third, if rack PDUs are basic and lack metering features, the facility lacks the visibility needed for effective capacity planning and efficiency improvements. Fourth, if structural integrity, door security, or physical access controls fail to meet current compliance standards—common in data centres over ten years old without rack refresh—the risk justifies replacement, regardless of efficiency gains.
For data centre applications, an online double-conversion UPS should have at least 96% efficiency at typical loads. Top-tier systems like the Galaxy VXL series can reach up to 99% efficiency in ECO mode without sacrificing protection standards. The efficiency you should focus on is the one at your actual operating load percentage, rather than the peak efficiency typically measured at 100% load—an operating point most data centres do not reach normally. Request efficiency curves for loads at 25%, 50%, 75%, and 100%, and compare them to your expected load profile. For a data centre with 500kW of IT load, the difference between a UPS running at 94% and one at 99% efficiency results in roughly 26kW of continuous energy loss, which costs around ₹18 lakh annually at standard Indian commercial rates.
Modern intelligent rack PDUs, InRow cooling units, and UPS systems from top vendors such as Schneider Electric's NetShelter and Galaxy series support common protocols such as SNMP, Modbus, and REST APIs. This enables seamless integration with most leading DCIM platforms regardless of vendor. EcoStruxure IT is designed as a vendor-neutral monitoring system, capable of collecting data from both Schneider and non-Schneider infrastructure in a single dashboard. For data centres with existing DCIM setups, this integration capability is crucial during procurement. Choosing cooling or PDU systems that rely on proprietary management platforms can increase licensing costs and lead to monitoring fragmentation, hampering operational efficiency. During the selection process, request a list of supported protocols and tested DCIM integrations, and verify compatibility with your specific DCIM version instead of relying on broad claims of compatibility.
Author Details:
Name - Anuj Kudesia
Email ID -Anuj.Kudesia@se.com
Company Name - Schneider Electric India
Linkedin Profile: https://www.linkedin.com/in/anujkudesia/
Author Bio: Anuj Kudesia, MD-SEPSL and Director of Business Development at Schneider Electric, leads strategy, product launches, and revenue operations for the Secure Power Division across the Greater India Zone. With over 20 years at Schneider Electric spanning enterprise data center sales, channel management, and strategic business development including P&L ownership for BFSI, IT/ITES, and Government verticals, he brings deep expertise in UPS systems, critical power infrastructure, edge computing, and data center modernisation for Indian enterprises.
Author Title: MD-SEPSL and Director of BD at Schneider Electric
ZZ_Paragraph Title
Need help?
Product Selector
Quickly and easily find the right products and accessories for your applications.
Get a Quote
Start your sales enquiry online and an expert will connect with you.
Where to buy?
Easily find the nearest Schneider Electric distributor in your location.
Help Centre
Find support resources for all your needs, in one place.