High-Speed Transmission Update: 200G/400G by sinovo telecom

sinovocorp

The IEEE 200G/400G standard is well on its way to publication; so where will it be implemented?

This is the first time in the development cycle of Ethernet standards that there have been so many different standards projects in the pipeline. Ranging from EPON over Coax (EPoC) to 200Gb/s and 400Gb/s Ethernet, these efforts demonstrate the continued varied application needs of end users. Let’s focus on the IEEE P802.3bs Task Force to see when the standard will be released, what types of products are being designed, and where they will be used.

IEEE P802.3bs Task Force

The IEEE P802.3bs Task Force was approved in March 2014 after almost two years as a study group. It started as 400G only, but in May 2016, 200G was added to the project. The scope now includes 200Gb/s over single-mode fiber and 400Gb/s over optical physical media. Optical and electrical signaling of 50Gb/s are also being developed to support both 200G and 400G transmission rates and are being standardized in the IEEE P802.3cd Task Force.

The current standards timeline is shown below.

200Gb/s and 400Gb/s Standardization Timeline (Source: IEEE P802.3bs Task Force)

The “bs” standard’s variants include the following:

• 200Gb/s
  • At least 500m of four-lane parallel SMF
  • At least 2km of SMF
  • At least 10km of SMF
    
• 400 Gb/s
  • At least 100m of MMF
  • At least 500m of SMF
  • At least 2km of SMF
  • At leas 10km of SMF
    
  
  

200Gb/s and 400Gb/s Applications

Internet service providers and network operators have already started to adopt 200G for both metro and long-haul networks. Most of these have been proprietary solutions from the likes of Huawei, Nokia, Ciena, Coriant, and Infinera. By 2018, when the Ethernet standard is released, we expect at least some of these to move to standardized solutions.

Cloud service providers like Amazon, Facebook, Google, and Microsoft are currently implementing 100G solutions both inside and between their data centers. These companies have plans to implement 200G and 400G within the next five years.

Connectivity to Support 200G and Beyond

While the transmission rates developed in the “bs” standard are 200G and 400G, the underlying signaling rates will range from 10G to 50G and perhaps ultimately land on 100G. Connectors, cables, and supporting hardware have been and will continue to be developed to support all of these solutions. The largest opportunities, by far, are in the data center market. The diagram below shows 50G signaling implementations for a top-of-rack (ToR) data center switch.

Top-of-rack (ToR) switch-to-server 200G Connectivity (Source: IEEE 02.3bs_CFI_01_1115.pdf)

Connectors and cables used in this solution are expected to be the following:

• SFP56 for single-port 50G ports
• QSFP56 for quad 50G ports
• QSFP56 for single 200G uplink
• Direct-attach copper (DAC) twinax cables to 3m to servers
• MMF or SMF modules or active optical cables (AOCs) for 200G uplinks
  

Clearly, using 50G signaling to achieve 200G and 400G transmission rates allows for flexibility in data center architecture, which is why we expect rapid adoption by service providers once the standard is complete.