Multi-fiber, Very Small Form Factor (VSFF) connectivity solutions are now required in areas previously serviced by MPO and LC based cabling infrastructure.   Conventional data center computing has been supported by networks with optical fiber links between the switch layers and copper or active optical cables at the server racks. As bandwidth requirements increase, and back-end clusters for AI computing grow, the number of optical ports are increasing in the switching layers and also now extend from the switch fabric to each server, increasing the fiber counts to be deployed both between racks and within racks in a data center.

Optical fiber cable assemblies traditionally terminated with MPO or LC connectors to service the connectivity within and between data halls does not achieve the density now required in the data center space. Structured cabling using MMC-terminated assemblies enables more optical port terminations per patch panel and rack, reducing space required for passive optical infrastructure.

It is increasingly common in large data center campuses to see multiple high fiber count optical cables, up to 6,912 fibers each, deployed between data center buildings for fabric aggregation layer links.  Traditionally, these cables are pulled or jetted through conduit and then spliced to multiple lower fiber count pigtails to create the optical link.   The same is true for constrained space pathways inside a data center, for example cabling installations between data halls or between suites in a multi-tenant data center.

With the critical operations running in a data center, speed of deployment and uptime are critical. With a reduced size compared to connectors such as the MPO, the MMC connector platform can be used to create factory-terminated assemblies that can be installed in constrained space pathways where previously only bulk cable and field fusion splicing was possible, enabling increased deployment speeds over traditional installation methods and reducing downtime of the data center.

As optical port count requirements continue to increase, driving density challenges within the structured cabling and passive racks in the data center, the same challenges are faced at the active equipment.  The size of MPO and LC connectors at dual-port transceiver interfaces can lead to difficulties with space and accessibility of the patching at the switch and server.

With a reduced size footprint, MMC connectors occupy xx% less space than the MPO  connectors in a dual-port transceiver, improving connector and patch cord accessibility at the transceiver interface as well as enabling increased areas for airflow at the active equipment. 

As data center link bandwidths increase and power efficiency improvements are needed, advanced optical interconnect technologies such as Co-Packaged Optics (CPO) and Near-Packaged Optics (NPO) may be necessary. Unlike traditional pluggable transceivers, these technologies relocate optical engines from the card edge closer to switch ASICs or GPUs, with optical fibers creating the link between the engines and the card edge. The optical engines, which may require a large number of optical fibers, are small in size, resulting in limited shoreline for optical fiber connectivity.​

The MMC platform supports the density required to enable CPO and NPO architectures with reduced size packages in the MMC Jr and MMC mini connectors.​

The optical fibers may be permanently attached or may be optically separable at the optical engine. As the fibers route on the board to the card edge, high density mid-board connectivity can be achieved with MMC. At the card edge, MMC high density solutions enable reduced space for the optical I/O fibers, allowing room for other components as well as more airflow.​