Lessons from 400G ease 800G data center deployments

Developments for 400G and lessons learned from data center deployments have put the industry in a great position to deliver 800G solutions and help service providers build out network capacity. #sponsored

Simon Stanley

December 20, 2022

3 Min Read
Lessons from 400G ease 800G data center deployments

A staged upgrade to 800G made possible by developments for 400G enables service providers to quickly deliver higher bandwidths and network capacity in data centers. Many service providers successfully migrated data centers to 400G by taking advantage of double-density QSFP-DD optical modules that are backward-compatible with 100G QSFP and by using optical breakouts to support four 100G connections per switch port. The same approach can be used for the migration to 800G; service providers can take advantage of the newest switches and 800G client optics without waiting for a complete 800G ecosystem to develop.

The deployment of 400G by hyperscale data center operators reflects the ever growing use of cloud services and the need for higher bandwidths to support demanding applications, including artificial intelligence (AI), machine learning (ML), and video processing. The introduction of 1RU switches with 32 highly flexible 400G ports gave service providers the opportunity to deploy more 100G connections using breakouts or upgrade connections to 400G. Backward compatibility with existing 100G solutions has been an important consideration in these deployments.

The shift to PAM4 modulation for both optical connections and electrical links between switch devices and optical modules and the introduction of double-density optical modules have been key to meeting the timeline needed by service providers for 400G optics. PAM4 modulation doubles the data rate per lane, and double-density optical modules double the number of lanes per module. Taken together, they quadrupled the effective data rate, allowing first-generation 400G optical modules and 12.8T switches to use silicon and optical technologies already deployed for 100G in data centers.

The introduction of 100G PAM4 lanes and the widespread use of silicon photonics, which are both critical for reducing the cost of 400G deployments, have enabled the rapid introduction of 800G solutions.

The first 800G ports using double-density modules have eight 100G PAM4 electrical lanes connecting to the switch silicon. The use of silicon photonics reduces the complexity and manufacturing cost of optical modules, especially those with many lanes. The benefits of silicon photonics will be increased further by the likely need for more lanes to support 1.6T and 3.2T solutions in the future and the ongoing development of co-packaged optics.

The backward compatibility of the double-density QSFP-DD form factor has given service providers additional flexibility as they have upgraded to switches with 400G ports, allowing them to manage the migration as demands on their networks have grown. This flexibility will also be important as service providers upgrade to 800G. The 800G QSFP-DD-800 form factor is compatible with 400G QSFP-DD modules, 200G QSFP56 modules and 100G QSFP28 modules. Modules are available for all these form factors that support port-to-port links, optical breakouts or direct attach copper (DAC) cables.

The recent introduction of 25.6T switches with 800G ports gives service providers the opportunity to start their 800G network upgrade to deliver higher bandwidths and network capacity in data centers. These 800G switch ports will already support two 400G connections, four 200G connections, eight 100G connections or a single 800G connection using dual 400G optical links. The 800G ports will also support a single 800G link once the specifications for the 800 gigabit Ethernet standard being developed by the IEEE802.3df Task Force are accepted and optical modules become available.

Developments for 400G and lessons learned from data center deployments have put the industry in a great position to deliver 800G solutions and help service providers build out network capacity. The use of backward-compatible double-density optical modules and breakouts gives end users the flexibility to manage the migration to 800G as demands on their network grow. This staged approach to network upgrades that worked well for 400G is enabling service providers to upgrade their data center networks to 800G before the introduction of 800 gigabit Ethernet.

Looking for more information?

— Simon Stanley, Analyst at Large – Components & Subsystems, Heavy Reading

This blog is sponsored by Cisco.

About the Author(s)

Simon Stanley

Simon Stanley is Founder and Principal Consultant at Earlswood Marketing Ltd., an independent market analyst and consulting company based in the U.K. His work has included investment due diligence, market analysis for investors, and business/product strategy for semiconductor companies. Simon has written extensively for Heavy Reading and Light Reading. His reports and Webinars cover a variety of communications-related subjects, including LTE, Policy Management, SDN/NFV, IMS, ATCA, 100/400G optical components, multicore processors, switch chipsets, network processors, and optical transport. He has also run several Light Reading events covering Next Generation network components and ATCA.

Prior to founding Earlswood Marketing, Simon spent more than 15 years in product marketing and business management. He has held senior positions with Fujitsu, National Semiconductor, and U.K. startup ClearSpeed, covering networking, personal systems, and graphics in Europe, North America, and Japan. Simon has spent over 30 years in the electronics industry, including several years designing CPU-based systems, before moving into semiconductor marketing. In 1983, Stanley earned a Bachelor's in Electronic and Electrical Engineering from Brunel University, London.

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like