Methodology for Implementing Modular Data Centers and Interpreting the Standardization Process
This article presents a methodology for modular data center implementation, emphasizing the essential role of standardization—distinguishing design from prefabrication—and analyzes how leading companies like Google, Microsoft, and Facebook apply these processes, while also comparing colocation providers such as Digital Realty Trust.
Before starting this chapter, the author poses four questions for readers to consider: when to involve equipment suppliers and integration solution providers, why major cloud providers adopt different modular designs, how to assess the depth of modularity, and which system to start with when implementing a modular data center.
The author introduces the crucial concept of standardization, arguing that modularity must be repeatable and scalable, which can only be achieved through standardized designs and prefabrication.
Standardization is divided into two layers: standardized design (the blueprint) and standardized prefabrication (the manufactured component). Consulting design firms provide the design, while equipment suppliers and integration solution providers deliver the prefabricated products that turn the blueprint into reality.
An everyday analogy compares standardized design to a doctor's prescription and standardized prefabrication to the manufactured pill that follows that prescription. Both are required for true modularity.
The "micro‑module" approach in China is examined; it typically achieves only standardized design, not prefabrication, unless additional on‑site work is performed. In contrast, container data centers are true examples of standardized prefabrication, as shown in the image below:
By classifying data‑center equipment into ICT and facility categories and mapping them onto the two standardization layers, a matrix (illustrated in the following diagram) can be created to visualize the overall standardization process:
Combining the axes of equipment type and standardization level yields the standardization progress diagram (shown below), demonstrating that modularity is the outcome of standardization:
Section 8 – Interpreting the Data‑Center Standardization Process
The remaining three questions from Chapter 7 are answered by applying the standardization framework to various companies. Google’s evolution is illustrated with timeline images, showing that Google covers all standardization stages and thus has a high degree of modularity:
Microsoft’s data‑center development is similarly documented with a timeline image, and additional details can be found in the linked video:
Facebook’s progress is shown with another timeline, confirming a high standardization level:
To contrast with internet giants, the author examines a colocation provider, Digital Realty Trust (DRT). DRT’s POD architecture has evolved to version 3.0, focusing on mechanical‑electrical modules, as illustrated by the images below:
Applying the standardization matrix to DRT reveals a lower overall modularity compared to the major cloud providers, primarily because colocation operators lack control over ICT customization and instead prioritize flexibility to accommodate diverse customer equipment.
The article concludes that while internet companies achieve high modularity through in‑house ICT design, colocation providers focus on mechanical and building modules, where higher standardization yields greater cost savings and performance benefits.
Alibaba Cloud Infrastructure
For uninterrupted computing services
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.