When people think about data centers, the first things that come to mind are probably exciting technologies like high-density server racks and liquid cooling solutions. Next in line might be a bit of the human touch, such as remote hands services or on-site security solutions. What most people probably don’t think about is cabling, despite the fact that effective cabling strategies can mean the difference between a high-performing data center and a facility that struggles with server downtime.
Fortunately, having good cabling strategies in place is a cost-effective way for data centers to improve performance and uptime reliability. When considering all the investment in a facility’s infrastructure, it doesn’t make much sense to leave the quality of its cabling up to chance.
First and foremost, every data center should be following a structured cabling plan regardless of deployment needs. While a small, on-premises data room may be able to manage a handful of cables connecting a few pieces of hardware without much difficulty, a data center environment is more complex by orders of magnitude. A poorly-managed cable distribution design and implementation process can cause complex problems for the future management of the data center, such as impeded air flow, difficulty with end-to-end identification, and removal of disconnected services which add time and complexity to the repair process. Without a formally managed plan, problems can occur and the resulting corrective actions are expensive in terms of cost and customer service.
A structured data center cabling solution that utilizes centralized patch panels minimizes the need for additional direct cable application, thus reducing the likelihood of cables being installed incorrectly. Today's data centers are seeing an increase of direct route circuits to meet the requirements for maximum bandwidth. Structured cable interconnection systems offer the highest quality of the initial installation or expansion, but a mandatory connectivity program that’s effectively maintained and managed is still and will always be an intricate part of the data center regime.
The next opportunity for data center connection efficiency is pre-terminated cabling. This type of cable comes from the factory with the connectors already installed. Using the correct pre-terminated cables provides an efficient “plug-and-play” solution for networking equipment. Historically, they have been used for “patch-cabling” where quick connections during a deployment are required and typically between two structured installations. However, recent applications have been developed for customer-to-customer connections or customer-to-provider connections.
Once the proper pre-terminated cable has been installed, the data center can maintain a cataloged inventory of applicable cables for similar installs or repair/restoration requirements, which reduces the service impact time for the network or specific customer. Pre-terminated cables are tested and certified by the production company, reducing the onsite time to install. It is critical to use appropriate lengths for permanent installation as excess cabling will result in storage and service issues over time. The result will be the same as mentioned above, airflow impedance, validation, and quality control.
A data center should use the highest grade copper or fiber cable to meet today's specifications and those of the future. Older fiber cables won’t tolerate the reduced bend radius required in today’s robust data center, nor will older versions of CaT cabling or twisted pair. The economics of using older cabling is alluring, but it will wreak havoc and result in an intolerable amount of severe issues.
Unfortunately, space is usually at a premium, requiring greater control of both quality material used as well as quality installed cable. Reusing cables left in place is tempting, but the time and cost savings will be forgotten when the service issues start. Whatever the cause, using older, degraded cabling can significantly affect performance, reducing bandwidth and even causing downtime due to dropped connections.
As a general data center cabling rule, use systems with automated production when possible. Never be totally reliant on a single source or method as flexibility is the key to meeting all of your customer expectations. Finally, all cost control decisions must be balanced to assess flexibility and quality of service.
Data centers use a lot of cabling to deliver services. Without proper documentation for server cable management, it’s easy to lose track of what types of cables are used in different deployments and when they were installed. Proper and thorough documentation makes it easier to maintain existing deployments and plan for new ones. If a data center is using pre-terminated cables for a network rack layout, managers need to know how many such cables will be needed to perform speedy repairs and connect new servers and equipment. Without documentation, technicians will waste valuable time assessing their cabling needs for every project. Having this information readily available allows them to handle problems and installations faster and with significantly less waste.
Cabling may not seem as important or as impressive as big-ticket items like power and cooling infrastructure, but it is no less critical to a data center’s performance. By following smart cabling strategies, data center operations can run more smoothly, allowing remote hands and migration personnel to spend more time delivering value to customers and less time tying themselves in knots over the facility’s cabling needs.
Ross is a Regional Vice President, Operations at vXchnge and is responsible for managing all 14 data center locations. With more than 30 years of experience, Ross has managed data center construction, engineering, repair and maintenance, leading him to the emerging business of colocation.