IEEE to consider what comes after 400 Gigabit Ethernet

April 23, 2020
The IEEE 802.3 New Ethernet Applications Ad Hoc will begin work at the end of this month toward a Call for Interest (CFI) to determine what transmission rate or rates the IEEE should consider next for possible standards work.

The IEEE 802.3 New Ethernet Applications Ad Hoc will begin work at the end of this month toward a Call for Interest (CFI) to determine what transmission rate or rates the IEEE should consider next for possible standards work. The action comes after the IEEE 802.3 Ethernet Working Group April 3 released IEEE 802.3™ Industry Connections Ethernet Bandwidth Assessment Part II. The study indicates that many of the applications to which Ethernet could be applied could see more rapid bandwidth growth rates through 2025 than the data rate increase a 1.6 Terabit Ethernet specification would enable versus 400 Gigabit Ethernet – assuming work on such a standard was completed by that year.

The bandwidth assessment is the second the IEEE 802.3 Ethernet Working Group has completed. The first, published in 2012, provided the catalyst for creation of the IEEE 802.3bs 400 Gigabit Ethernet standard (see “IEEE 802.3 400 Gigabit Ethernet Study Group chair warms to task”). Work on the latest study began in 2018, shortly after completion of the IEEE 802.3bs efforts (see “IEEE to assess Ethernet bandwidth demand”). The assessment report is available from the IEEE 802 website.

Several aspects of the assessment report stand out, according to John D’Ambrosia, senior principal engineer, IP Standards Team at FutureWei Technologies, leader of the bandwidth assessment effort, and editor of the report. One is the dynamic characteristics of the various factors that determine bandwidth demand – the number of users, access rates, access methods, and applications all contribute in ways that can easily change over time. And those applications of Ethernet are growing in number. The study examines such applications as data center capacity, global content distribution networks (CDNs), global IP, IP video, mobile, peering, and virtual and augmented reality (VR/AR). D’Ambrosia points out that artificial intelligence and connected cars (data for which was not available when the assessment report was written) also should be considered as well in pondering likely future Ethernet applications that could increase Ethernet transmission rate demands.

If one puts the amount of traffic each of these applications generate now and how much each is expected to contribute over the next five years, “You get some really scary, scary numbers,” D’Ambrosia said. That’s why he issued a request within the Ad Hoc to begin development of the CFI which, when launched, will help 802.3 achieve a consensus on whether new transmission standards will be necessary in the near future and what those rates might be.

The Ethernet Alliance, the industry association dedicated to the advancement of Ethernet technologies, also has plans to help drive consensus on the topic. The group has planned a Technology Exploration Forum (TEF) on “The Next Ethernet Rate” for October 7, 2020, at the Santa Clara Convention Center, Santa Clara, CA. (The event was originally scheduled to take place this month but was rescheduled due to the effects of the COVID-19 coronavirus pandemic.) The event will feature a series of panel discussions, presentations, and discussion opportunities on what should come after 400 Gigabit Ethernet. More information and registration details are available on the Ethernet Alliance website.

Where will Ethernet go next?

So what options are likely to be debated within the CFI activity? The bandwidth assessment report considers 800 Gigabit Ethernet and 1.6 Terabit Ethernet in its commentary. D’Ambrosia notes that among the considerations for a new specification is whether it will result in something that will meet industry expectations for cost, power, performance, and availability of enabling technology. With the near-term availability of 100-Gbps electrical lane specifications that would match the achievement of single-lambda 100-Gbps optical technology, both 800G and 1.6T should be technically feasible as 8-lane and 16-lane schemes, respectively, D’Ambrosia said. Despite the release of an 800G Ethernet specification from the Ethernet Technology Consortium (see “25 Gigabit Ethernet Consortium renames to Ethernet Technology Consortium, launches 800GBASE-R 800GbE specification”), he adds a formal IEEE 800 Gigabit Ethernet standard is desirable.

Nathan Tracy, Ethernet Alliance board member and technologist within TE Connectivity’s Data and Devices business unit, agrees that 800 Gigabit Ethernet would be a natural next steppingstone – but some of the major data center operators will want something faster, such as 1.6 Tbps. Both Tracy and D’Ambrosia stated they would not be surprised if some parties lobbied for something at a higher rate than 1.6 Tbps. But they pointed out that an efficient, economical technological pathway to such a higher rate isn’t as clear as that toward 1.6 Tbps.

For related articles, visit the Optical Technologies Topic Center.

For more information on optical modules and suppliers, visit the Lightwave Buyer’s Guide.

To stay abreast of optical communications technologies, subscribe to Lightwave’s Enabling Technologies Newsletter.