Shanghai Huijue Network Communication Equipment Co., Ltd
sales@hj-net.com
021-57471000
13916913483
Share the method of increasing the port density in the distribution frame in the data center
column:industry news Release time:2022-10-06

The demand for network connectivity in data centers is on the rise, especially in hyperscale data centers that increasingly use 17,286-fiber or 34,566-fiber cables. Connecting this amount of fiber to servers and switches is a key challenge because rack space is limited. Fiber distribution frames are at the heart of solving this challenge. In order to solve this problem, industry manufacturers are increasing the port density in the patch panel to meet the ever-increasing bandwidth demand.

 

 

Maximize rack space

 

One of the challenges data center architects face is maximizing the use of valuable rack space. Each rack typically has 42 rack units (RU) or 48 rack units (RU), and the main goal is to use that space to deploy as many servers as possible. Obviously, the less space you have for passive components, the more space you have for active hardware. Data center architects want to maximize the use of valuable data center real estate to deploy network connectivity equipment, and increasing port density is an effective solution.

 

Meanwhile, when it comes to network connectivity, data center architects employ different application and fiber optic interfaces. Some use serial connections, while others use parallel connections to increase bandwidth. Ideally, passive infrastructure and horizontal cabling provide as much flexibility and modularity as possible, avoiding the need to purchase different types of wiring solutions for different needs. Network connector formats are also evolving: simplex SC and duplex LC have been popular over the years. Alternative miniaturized serial connector formats (such as Senko’s SN, USConec’s MDC) are participating in the latest “connector wars.” In the past ten years, parallel connection technology has made great progress.

 

But the performance of network connectors is constantly improving, and everyone wants to have the most flexibility when choosing the type of network connector. Whether it is a duplex connector with two fibers per port or a parallel connector with a greater connection range per port. USConec has certified 72 fibers per port with MTP, while 24 fibers per MPO/MTP port are common. Port preference can move from serial to parallel, and eventually serial technology. Without a flexible platform, the number of future-proof MACs will be limited and the time horizon of the investment will be limited.

 

Patch panels are where connections are made between computing hardware and high-fiber-count cables. Users will need to deploy shielded cables and latches for the ports on the patch panel during initial deployment. Users need good cable management (conceived at platform design time and deployed at installation time) and access to port latches. In some fixed panel designs, the need for access to the front and rear of the panel affects flexibility.

 

Some vendors offer panels with adapters that may provide more ports per panel, but they ignore the cable management aspect and don’t offer flexibility or modularity. Removable panels provide better port latch access, but tend to fall short in terms of modularity and flexibility.

 

Most of these solutions offer limited modularity, so users must be aware of port types when purchasing panels. Also, when filling the ports of a fixed panel at maximum density, there is no panel space available for port identification. When using fixed panels, in order to obtain maximum density, the panel design may deviate from the space of 1RU panel, and must adopt 2RU, 3RU or even more rack space to achieve the highest port density.

 

Traditional modular patch panel designs limit some manufacturers to 72 ports per rack unit (RU). Some vendors have non-interchangeable platforms that are not flexible. Other solutions sacrificed port density to maintain interfaces to automated infrastructure management (AIM) solutions, and this backward technology has prevented many vendors from increasing fiber density. Additionally, many vendors do not want to introduce new platforms that are not compatible with previous platforms.

 

Flexibility and modularity are key

 

The value of flexibility and modularity is that the platform can be reconfigured to suit different needs for delivering different applications. If the platform is modular at the connector level, the serial and parallel ports can be accommodated in the same termination box if desired. In addition, modular connectors allow port assignments through the use of colored adapters, keyed adapters, bezeled and unbezeled adapters, and more.

 

Platform-level modular equipment also allows customers to use the same termination box in multiple applications while maintaining port density. If the user uses a cable with a high core count, pre-termination at both ends may not be possible due to limited paths (such as small conduits). The modularity of the platform should extend to termination boxes, so the same termination box with different internal components can accommodate its internal or external field splicing.

 

Today, MPO has become the preferred parallel optical connector for data centers, mainly using 12-core optical fiber. Recently, 8 active fibers in a 12-fiber ferrule were found to be favored based on breakthrough applications. And its platform should at least support these options. The ideal solution would support both 8-fiber and 12-fiber connections, with additional features to support newer parallel connection variants that include one row of 16-fiber and two rows in the same MPO, 24-fiber and 32-fiber fiber optic connection. Ideally, patching platforms will be able to mix patch boxes within the same enclosure without loss of port density. Most solutions support two or three of these.

 

There will be a trend towards more use of parallel connections in the future, which is usually the first step in increasing bandwidth. As transceiver processing speeds increase, serial applications are often first to market. By adding more lanes (fibres), bandwidth can be increased at the same processing speed. Periodically, as transceiver efficiency and speed improve, serial connection solutions tend to replace earlier parallel solution connections. Most data centers start with serial applications, increase bandwidth as parallel applications, and return to serial connection solutions after a few years due to improved transceiver efficiency. Termination platforms should have the flexibility to easily migrate connections from serial to parallel and back to serial as the application grows.

 

An ideal connectivity solution should accommodate these trends, easily support serial and parallel connections, and accommodate different types of connectors and cable management requirements. By designing patch panels with a high degree of modularity and flexibility, fiber optic suppliers can provide more support for data center architecture.

TOP