Data center design standards, such as the Uptime Institute’s Tier Standard and TIA-942, place quite a bit of emphasis on the power and cooling systems, which is justified as these play major roles in the total operating cost and overall functionality of the data center.
However, an equally important but sometimes neglected design aspect is that of the physical infrastructure required to house and interconnect all the active equipment. The primary components of the physical infrastructure are the racks/cabinets, the structured cabling, and cable management, which are integral to the overall efficiency, functionality, and scalability of the data center regardless of its size, purpose, or complexity.
By following some best practices for data center infrastructure design a lot can be done to help increase efficiency, improve functionality, prepare for future growth, and minimize unnecessary or unplanned downtime.
Data Center Design Best Practice #1: Build with an Eye Toward Growth and Scalability
In today’s business environment with a future that seems to be evolving into the Internet-of-Everything, the demands on the data center are ever increasing, regardless of how large or small. Because floor space is always at a premium, it is not cost-effective to incorporate into the design a significant amount of unused floor space dedicated to future growth. However, it is possible to optimize the use of available space by installing racks/cabinets with some or all of the following characteristics:
- A modular design that is able to be moved, installed, adjusted or otherwise modified with relative ease
- Higher weight limits to support new equipment
- Taller heights to accommodate vertical growth
- Adjustable front and rear mounting rails to accommodate multiple equipment depths
- Able to support different varieties of cable management and able to change and be made available wherever required
- Integrates easily with overhead pathways
Data Center Design Best Practice #2: Consider the Quality of the Cable and Connecting Hardware
An often-overlooked aspect of data center build outs is the need for a high-quality cabling system. Low-quality cabling components combined with poor installation practices can negatively impact the overall performance and reliability of the network, which could lead to costly, unplanned outages.
Be sure to only install cable and connecting hardware from a reputable manufacturer that offers extended warranties as well as initial design support.
Data Center Design Best Practice #3: Place Emphasis on the Cabling Architecture
The active equipment within a data center is usually distributed among multiple racks/cabinets and often further segmented by function, department, business group, etc. Therefore, it is imperative that consideration be given to how all this equipment will be connected. It is not enough to simply count the number of connections required in each rack/cabinet and assume that providing the same number of copper and/or fiber cables will suffice.
The more common topologies for the distribution of cables and connecting equipment are Top-of Rack (ToR), Middle-of-Row (MoR), and End-of-Row (EoR). These topologies can be used exclusively or together depending on the size, complexity, and specific requirements of the data center. However, they all have a significant impact on how and where cables should be installed as well as the size and type of cable management required in each cabinet and/or each row of cabinets.
Data Center Design Best Practice #4: Consider How Well the Cabling Infrastructure Can Accommodate Future Growth and Emerging Technologies
Networking speeds and bandwidth requirements have traditionally been on the rise and this trend is not likely to change. It is not uncommon for 1GBase-T and even 10GBase-T to be deployed out to the workstation, which places even more demand on the data center to support the increased bandwidths. The emergence and the increasing adoption of 40G and 100G have forced the adoption of MTP® connectors for optical fiber cables, which is a major departure from traditional fiber termination practices and often requires pre-terminated assemblies.
Although these networking speeds require optical fiber, copper cable is still very prevalent and continues to be a mainstay in data center design. Therefore, a hybrid infrastructure consisting of both copper and fiber cabling is highly recommended but the design must be based not only on the current needs of the network but also on future requirements and the ability of the system to adopt new and emerging technologies. A properly-designed system should be modular and able to easily accommodate additional cables, multiple types of media and different types of connections/terminations.
Data Center Design Best Practice #5: Pay Close Attention to Organization and Proper Documentation
As data centers grow and evolve over time, the cabling infrastructure will likely need to grow and evolve along with it. This is especially true if future growth was not sufficiently factored into the initial design.
One way to mitigate the effects of change is to implement and strictly adhere to a labeling scheme as well as a change management process for equipment upgrades, connections, and additions. The more organized and structured things are, the less likely things will get out of control and become dysfunctional.
If adequate cable management is not designed into the initial infrastructure and administration practices are not implemented or followed, then the cabling may quickly become a jumbled mess—kind of like Cousin Itt from “The Addams Family” on a bad hair day. When things get out of hand, it may be necessary to perform a data center cleanup to restore order, which usually involves downtime and can be somewhat costly.
Have any questions about best practices for data center infrastructure design? Reach out to the experts on the Blue Wave Communications team today!