The Top 5 MEP Challenges of Data Centers

Data centers are unique builds that come with unique challenges. MEP contractors stepping up to build data centers are sure to encounter the complexity of systems and the demands of scalability, redundancy, environmental control and IT infrastructure integration. 

This article explores the five key challenges that are essential to data center construction and demand close attention from MEP contractors.

Data Centers: Unique from the Start

Humans work at data centers, but the primary consumers of the completed structure are the servers occupying row after row of racks. From a few hundred to millions in one space, data center servers communicate among themselves to exponentially increase their processing power. 

The high demands of servers for precisely controlled operating conditions place pressure on construction. Data centers must be completed on strict deadlines, and they must be operational from Day One. 

Successful MEP contractors approach data center challenges with an understanding of the building’s purpose, power and cooling requirements and spatial constraints. 

With their insights, they ask questions and pursue follow-ups to understand the owner’s intent and long-term goals. They don’t see drawings as their marching orders for the day but as springboards for conversations about the future growth and needs of the data centers that are now central to the digital economy. 

Under these criteria, five particular challenges face MEP contractors working on data centers.

1. System Complexity

Data systems house multiple systems that must work in harmony. They don’t just cross paths, to be resolved with traditional clash detection techniques. They work together — for instance, plumbing and mechanical systems convening to support the center’s stringent cooling needs.

This level of integration requires advanced modeling and simulation tools to visualize interactions, prevent conflicts and maximize the completed center’s operating efficiencies. 

2. Scalability

Some data centers, including the hyperscale data centers built by the world’s tech giants, are slated for future growth. Owners know that demand will rise, especially as AI, with its hunger for processing power, consumes a rising share of digital interactions.

In a growth-oriented industry, scalability equips owners to leverage the sites they have already found and the structures they have already completed. By overbuilding, they bypass the time- and resource-consuming process of starting from scratch.

But as construction techniques and technology evolve and advance, no one knows what the future will look like. While planning some systems for scalability might not deviate from typical practices, electrical contractors should account for the IT load from the perspective of the future electrical requirement. 

This includes creating room for added server racks with their unique layout of separate aisles for cold-air intake and hot-air exhaust. A scalable mindset also considers the fact that future server racks might be designed differently than today’s or should be modularized to accommodate the unknown racking needs of the future.

3. Redundancy

While redundancy is common to many projects, it can be an outsized challenge for data centers. Any problems with power, cooling, vendors, or IT systems threaten downtime that could bring critical operations to a halt. 

Redundancy is so nuanced for data centers and other large, mission-critical projects such as hospitals that it is classified into four tiers. 

Data centers that can withstand a small amount of downtime are classified as Tier I, needing no redundancy, and Tier II, which might call for backup generators or extra uninterruptible power supply.

Tier III can only take 1.6 hours of downtime a year, requiring N+1 redundancy, which provides a single independent backup for every set of components needed for normal operations. 

The restrictive Tier 4 is imposed on data centers requiring 100 percent uptime or only fractionally less. These Tier 4 projects demand 2N redundancy, meaning that every system is fully duplicated and ready to run in case the main system fails. 

4. Environmental Controls

Servers require a staid and stable temperature strictly set at 68 to 72 degrees. The cooling system must operate continuously, with no breaks or fluctuations and minimal humidity.

With their rigid cooling requirements, server spaces must operate separately from human-occupied areas and their need for fresh-air makeup, which introduces humidity and temperature fluctuations that servers can’t tolerate. 

Even if placing people closer to the servers they monitor would simplify systems operations and be more convenient for data center employees, the two areas must be segregated.

5. IT Infrastructure Integration

Building what’s on the drawings will get the job done, but it won’t lead to long-term relationships with the companies that build data centers. They want partners who learn to interpret their assignments through the building’s IT needs, which are the reason for its existence. 

For instance, some data centers operate solely for the needs of the owner, such as a tech giant collecting, storing and processing data for internal use. There, running the servers as a neural network without interruption is top priority, so cooling is the primary consideration. 

However, a data center built to support e-commerce operations might prioritize a consistent and reliable network of fiber optic cables capable of carrying vast data loads in and out.

"We have real data with integrity to base our decisions on."

Learn how Vantage Data Centers created a single source of truth to improve data analytics, financial performance and communication across five continents.

Read Their Story

Overcoming MEP Hurdles in Data Centers

For MEP contractors pursuing data center construction, understanding the unique demands of these complex facilities is key.

To stay on the cutting edge of data center builds, contractors can prioritize advanced modeling and simulation tools to visualize system interactions and prevent conflicts, leading to more efficient operations.

It's also important to adopt a scalable mindset, accounting for future IT load and diverse rack designs, to accommodate the industry's rapid growth and technological shifts. This forward-thinking approach will allow contractors to deliver high-performing data centers that meet current demands and adapt to future advancements.

Written by