Data Center Cabling Designs and Significant Considerations of Its Operation

Data centers possess an immensely intricate cabling structure. A proper cabling design is integral to the functioning of the data center. A minute error in the data center cabling can lead to structural challenges such as spaghetti cabinets, ill-connection between network switches, and tedious product installation. The complexity of data center cable architecture eventually leads to problems in error-identification, maintenance, and troubleshooting. Adding to it, today’s modern technologies like the Internet of Things (IoT), cloud computing, big data, etc. demand optimum access to the highest storage capacity of the data center. Again any errors in data canter cable infrastructure can lead to operational challenges like data loss, insufficient storage, and abrupt breakdown of the data storage system. To avoid these risks, it is essential to integrate data center cabling in specific types of cabling that are called topologies while paying attention to a few significant considerations. This white paper guides readers through ways to perform effective and efficient data center cabling, its considerations, and performance-enhancing strategies.

General Considerations of Data Center Cabling

To perform efficient data center cabling in specific topologies, it is essential to consider a few significant factors. These factors add to the quality and efficiency of the data center cabling.

• Data Center Cabling Standards:Increasing data demands and shrinking loss budgets offer several challenges to data center engineers. However, gaining a fair understanding of infrastructure cabling standards may help them to minimize risks.
  • ANSI/TIA-942: This standard offers specific recommendations for efficient cabling in the data center architecture. These recommendations include a minimum number of pathways, space between horizontal cabling and backbone, environmental considerations, and guidelines for cable management, utility, and redundancy, etc.
  • ISO/IEC 24764: This standard is a combination of TIA-942 and EN- 50173-5. It describes the types of cabling suitable for specific data center requirements.
  • ANSI/BICSI 002-2014: This standard offers guidelines for specifics of the design and operation of data center cabling. It includes aspects like infrastructure design and planning, commissioning, maintenance, troubleshooting, etc. It covers minute aspects like pathways, spaces, infrastructure detailing, cabling classes, etc for different types of data centers.

• Up-Time and Security: It is essential to determine the data center security and uptime requirements of the overall data center. To increase the performance and life expectancy of the data center efficient establishment of cabling uplinks, power sources and cablings, etc are essential.

• Scalability: The network must be scalable. Therefore, it is important to calculate the expectancy of network performance, determine power and resource requirements according to calculated performance needs. The consideration of space between the servers and switches of different functional areas defines the scalability of the network.

• Manageability and Performance Flexibility: It is essential to have a manageable and flexible cabling network. Therefore, it is important to consider factors like configuration and reconfiguration frequency, cabling limitations, pathways and routing requirements, etc. The network must be functionally flexible to admit such changes after the first layout of the cabling infrastructure.

• Cost of Operation: The data center cabling includes several components like switches, connectors, media converters, etc along with the cables. Thus, capital investment, cost of operation, maintenance and damage coverage, etc. becomes essential considerations.

• Functional Areas of Data Center Cabling: The data center is distributed in different functional areas through which the cabling runs and adds value to the data center operation. To establish an efficiently designed data center infrastructure, the consideration of the following functional areas is suggested under the standard TIA-942.

• Entrance Room: The entrance room is the main interaction area of the data center from where the data center information can be accessed via an interface. This entrance room provides an access to operational equipment, demarcation points, and switch access to different functional areas.

• Main Distribution Area (MDA): The main distribution is area is where functional components are housed. It houses switches, routers, cross-connect, and inter-connect equipment, etc. The LAN essentials are housed in the MDA.
• Horizontal Distribution Area (HDA): HDA houses the connecting equipment and switches to interconnect equipment distribution area (EDA) to local area network (LAN), storage network area (SAN), and/or KVM switches.
• Zone Distribution Area (ZDA): ZDA is basically a coherent point between the HAD and EDAs.
• Equipment Distribution Area (EDA): EDA is the functional zone of the data center that houses end equipment like servers, racks, cabinets, etc. Also, the HAD is terminated in the EDA region by using patch panels.
• Backbone Cabling: Backbone cabling is the cable line that interconnects all other functional areas of the data center cabling. Error in backbone cabling can terminate the function of all other functional zones of data center cabling. That is why critical care of backbone cabling is essential to perform.

Taking all the above-mentioned factors into consideration, commonly three topologies of data center cabling are adopted across industries. These strategically planned topologies ensure lesser possibilities of errors, easy maintenance, and highly-efficient performance of the data center infrastructure.

Data Center Cabling Designs/Topologies

In order to successfully connect the data center cables, one of the following topologies can be utilized as per the performance requirements.

• Centralized Cross-Connect Topology: Centralized cross-connect data center cabling can be adopted in two ways:
  • Centralized Cross-Connect via MDA: In this type of topology, the backbone cable directly connects the switches from the main distribution area (MDA) to the horizontal distribution area (HDA), which is further centralized by cross-connecting the distribution rows or lines from HDA to the equipment distribution areas (EDAs). The distribution lines from HDA to the EDAs are connected by using three-connector channels, which allows the HDA lines to terminate at patch panels and replicate or mirror it to the switch ports of EDA. The termination point of the first patch panel is further cross-connected with the second patch panel so that each switch port from individual EDAs are cross-connected in the network.
  • Centralized Cross-Connect via ZDA: To implement this type of centralized cross-connect data center cabling, a four-connector channel is utilized. At first, the backbone cable connects the zone distribution area (ZDA) to the HDA. Further, by using a four-connector channel, the ZDA distribution lines are cross-connected to the patch panels of EDAs. However, the patch panels of the four-connector channels are interconnected with a cross-link and one or/and both links are cross-connected to the switch ports of different EDAs.
    Although centralized cross-connect via ZDA type of data center cabling is not widely adopted, it still can be used for data centers with critical security needs. It secures the cabling by efficient cooling and increases efficiency due to the easy reconfiguration feature.
• Benefits of Adopting Centralized Cross-Connect Topology
  
  The following are a few benefits of choosing a centralized cross-connect topology.
  • This infrastructure enables easy connections from any switch to any device in the data center cabling network.
  • Permanent connections to the switches can be established in order to avoid over-interaction with sensitive devices.
  • The cabling infrastructure enables easy configuration and reconfiguration, device addition, line integration, and so on.
  • Network flexibility and manageability are higher in these topologies.
  • The centralized cross-connect topology assures optimized space utilization which reserves server cabinet spaces.

• Partially Centralized Inter-Connect Topology: The partially centralized inter-connect topology of data center cabling features a backbone cabling from the main distribution area that runs through multiple HDAs in a row of specific EDA. While running through HDAs the backbone cable inter-connects the switches from each HDA with the help of two-connector channels. The second distribution lines from the two-connector channels are inter-connected to the patch panel of the EDA.

• This topology can be implemented in either of the following ways.
  • Point-to-Point: To implement the point-to-point approach to partially centralizing inter-connect topology, the switches from HDAs are allocated at a distance from the EDA. However, no cross-connect or inter-connect approach is deployed in this type. Long patch chords are used to connect the switches from HDAs with the patch panels in the EDA.
  • One-Connector Channel: To implement a one-connector channel approach, one-connector channels are used instead of two-connector channels. The patch panels are only deployed in the HDAs and then they are directly connected to different equipment housed in the EDA. This approach is suitable for smaller data center applications. It also allows complete utilization of mounting rail in the server cabinets.
• Benefits of Adopting Partially Centralized Inter-Connect Topology
  
  The benefits of partially centralized inter-connect topology are listed below.
  • It reduces the requirement of copper or optical cabling run.
  • It optimizes the pathway spaces and increases the safety of the connections.
  • The number of terminations is comparatively lesser than other types of cabling due to a lesser number of switches and patch panels being interconnected.
  • It allows the switches and patch panels to locate in the adjacent cabinets without jeopardizing the cabling efficiency.

• Distributed In-cabinet Switching: In the distributed in-cabinet switching topology of data center cabling, the connections are simplified by eliminating the use of HDAs totally. This topology features a backbone cable that connects the MDA directly to different cabinets or switches of EDAs. This way, the backbone is directly connected to performing devices in the data center connections.
  This type of cabling also features a direct connection to the storage area network (SAN) which further simplifies the data storage operations in the data center.

• Benefits of Adopting Distributed In-cabinet Switching Topology
  
  The following are the benefits of implementing distributed in-cabinet switching topology in the data center cabling.
  • This type of cabling drastically reduces cabling complexity and expenses.
  • For applications with high-bandwidth like 40gbps to 100gbps, this topology gives maximum returns on investment (ROI).
  • In special applications where a specific functional area of the data center is required to be separated or secured.
  • It is an effective solution for limited cabinet space availability.

Important Devices for Data Center Cabling

In addition to the aforementioned data center cabling topologies, it is also important to choose the right devices for building a robust data center cabling infrastructure. The following are a few important devices for consideration.

• Media Converters: Media conversion has emerged as a critical function in enterprise data centers. This is because most data centers utilize hybrid cabling infrastructure where copper cables and fiber optic cables are used together. These cable types have different media, which possess challenges when used together. This is where media converters help. The media converters enable data center managers to optimize the value of existing legacy infrastructure while taking advantage of fiber optic cables. Copper to fiber media converters and fiber to fiber media converters are two popular types of media converters used in data centers.
• Network Switches: The enterprise network infrastructure underwent a great change in the decade 2010-2020. With increasing emphasis on network-function virtualization (NFV), software-defined networking (SDN), and software-defined WAN (SD-WAN), choosing the right network switches have become important for data center engineers. You can see different types of data center-class network switches in the market. PoE/PoE+ Switches are also available with flexible ports in managed and unmanaged versions.
• Optical Transceivers: Until a decade ago, fiber optic technology was only used to mitigate critical interconnection challenges. However, today, it has evolved as a critical element of data centers. The fiber optic cables and devices assure better scalability, security, and performance, which makes them an integral part of data center cabling. Optical transceivers with SFP modules are used to extend distances, reduce power consumption, optimize cabling density, and leverage existing multimode fiber. Owing to all these reasons choosing the right optical transceivers is important. Today, these transceivers are available in various specifications such as 10/100/1000Mbps, 1000 Mbps, 100 Mbps, and 10 Gbps speeds.

Why do we need Networking Switches to build Data Center

Networking switches are essential components in building a data center due to their crucial role in managing and optimizing network traffic. Here are some key reasons why networking switches are necessary for data centers:

• Traffic Segmentation: Data centers handle a massive amount of network traffic from various sources, including servers, storage devices, and users. Networking switches enable the segmentation of this traffic into different virtual LANs (VLANs), ensuring that data flows efficiently and securely within isolated segments.
• High Performance: Switches are designed to handle high data rates and low latency, allowing data to be transmitted swiftly and efficiently between devices within the data center. This is crucial for maintaining the responsiveness of applications and services.
• Scalability: Data centers can consist of hundreds or thousands of interconnected devices. Switches provide the necessary scalability by connecting numerous devices while maintaining consistent performance and minimizing congestion.
• Load Balancing: Switches can distribute network traffic across multiple paths, preventing bottlenecks and ensuring that the load is balanced evenly across the network infrastructure. This results in optimized performance and resource utilization.
• Redundancy and High Availability: Redundant networking paths and failover mechanisms are crucial in data centers to ensure uninterrupted operations. Switches can be configured to provide redundant connections, allowing traffic to be rerouted in case of link failures.
• Quality of Service (QoS): Data centers often host a variety of applications with different network requirements. Switches can prioritize and manage traffic based on QoS settings, ensuring that critical applications receive the necessary bandwidth and low latency while preventing less critical traffic from consuming excessive resources.
• Security: Switches enable the implementation of security measures such as access control lists (ACLs) and port security. These features help protect the data center's infrastructure and sensitive information by regulating who can access the network and which devices they can communicate with.
• Centralized Management: Modern networking switches come with management interfaces that allow administrators to configure and monitor the network from a centralized location. This simplifies the management and troubleshooting of the data center's network infrastructure.
• Network Convergence: Data centers often use a variety of traffic types, including data, voice, and video. Networking switches support network convergence by allowing different types of traffic to share the same physical infrastructure while maintaining their specific quality requirements.
• Flexibility and Adaptability: As the data center's requirements evolve, networking switches can be reconfigured, upgraded, or replaced to accommodate changing needs without requiring a complete overhaul of the network architecture.

Guidelines to Improve the Performance of Data Center Cabling

In order to enhance the efficiency of a data center, several tactics are being utilized. Implementing fiber optic cables instead of copper cables for long run cross or interconnections. Even if the data center features legacy devices that are not compatible with direct fiber optics connections, the copper cables from the device switches can be connected to fiber optics cables to get higher data transmission speed, bandwidth, and data security. It is achieved by using devices like media converters, fiber optics switches, fiber optics connectors, etc. as copper-to-fiber moderator equipment. Along with replacing copper cables with fiber optics, there a few tactics that can help in enhancing the overall efficiency of the data center cabling. The guidelines to adopt the tactics are listed below.

• Utilize pre-terminated cabling to perform the connections. It enables plug-and-play operation which saves efforts and time on cabling integrations.
• Use high-quality interconnecting products like media converters, optical transceivers, networking switches, etc.
• Adopt flexible yet structured types of cabling topologies.
• Utilize back to back vertical cable managers to secure cable bulking at high-density locations.
• Form patching zones between the equipment and cable distribution lines in order to identify and categorize the elements.
• Prevent cable slack by using Velcro ties and guide the patch chords thoroughly.

By following the above-listed topologies and guidelines, it is eventually possible to build a high-performance efficient, cost-efficient, manageable, and durable data center cabling structure.

Conclusion

Considering the complexity and importance of different elements of data center cabling, the factors that need meticulous attention during data center infrastructure formation are listed in this white paper. In order to provide, global industrial compatibility of data center cabling, three main industrial standards are described followed by different approaches to designing a data center cabling network. Centralized cross-connect cabling, partially centralized inter-connect cabling, and distributed in-cabinet switching type of cabling topologies are described with their potential approaches and benefits. Besides this, a few important devices for data center cabling infrastructure and guidelines to improve the overall efficiency of a data center are described in this white paper. You can contact us to know more about the different solutions offered by the company to build an effective data center cabling infrastructure.