As 2023 continues to unfold before us, we have been thrilled to celebrate a significant landmark - the golden anniversary of Ethernet. Over recent months, we've been crafting and sharing an assortment of intriguing and enlightening content to honor this occasion. Our intention for this series has been to illuminate the multifaceted evolution of Ethernet and to recognize its critical role in shaping our globally networked society. Remarkably, even after half a century, Ethernet technology remains a silent yet significant force driving our interconnected world
Our most recent contribution to this series was a specially curated video titled, "Celebrating 50 Years of Ethernet - The Past, Present and Future," released on May 22nd, 2023, the very day of Ethernet's 50th birthday. Prior to the video, we enriched our series with an eye-catching infographic, offering a concise yet comprehensive look at Ethernet's historic milestones and its future potential.
The series was initially set in motion with my introductory blog that ventured into the origins and early development of Ethernet, harking back to its birth in 1973. The blog presented a chronology of Ethernet's journey and the pivotal milestones that have sculpted its path. This narrative was further enriched in a subsequent blog post, which offered a deep dive into the myriad of innovations sparked by Ethernet throughout its half-century lifespan. As we wrap up this content series, I will delve into the possible future capabilities of Ethernet which will continue to help shape our connected world.
our connected word
In my previous blog post, I explored how Ethernet has found its place in mission critical applications such as industrial automation, 4G and 5G mobile networks, and even in the most demanding rugged environments. I also touched on the concept of Time Sensitive Networking (TSN), a significant enhancement that allows Ethernet to evolve into a deterministic networking technology. TSN empowers the precise synchronization of network elements and endpoints, which facilitates traffic class prioritization, provides predictable latency, and guarantees bandwidth reservation.
One might contend that Ethernet is one of the most crucial technologies today, even though it often goes unnoticed. This can be primarily attributed to its ubiquitous presence, as Ethernet deployments stretch across the cosmos, extending from the deepest ocean trenches to the boundless expanse of outer space. The applications of Ethernet continue to evolve and broaden, including use cases for faster data transmission over greater distances. For instance, submarine cables that crisscross the ocean floors serve as intercontinental transport systems, further highlighting Ethernet’s expansive reach and continued growth.
As we’re , there’s an increasing number of applications being developed that have substantial bandwidth requirements, such as high-definition 8k video. This demand necessitates the underlaying transport technologies provide greater bandwidth. Consequently, 400G Ethernet is reality today, and 800G Ethernet is expected to become commonplace in the near future. Given this trend, it wouldn’t be surprising to see 1 terabyte Ethernet in use by 2030, as our appetite for bandwidth shows no signs of diminishing.
The emergence of machine learning (ML) and artificial intelligence (AI) is escalating the bandwidth requirements on hyperscalers like Microsoft, Amazon, and Google, pushing their data center infrastructures towards adopting 100G Ethernet to handle this increasing capacity demand. Intriguingly, this trend of increasing bandwidth demand is not exclusive to hyperscalers’ data centers but also evident with service providers’ transport networks, as both are adopting parallel strategies to manage this growth.
Within the Ethernet Alliance, there are a number of interesting new developments surrounding higher speeds such as 100G and 400G operation over single-mode fiber at 100G per wavelength. This standard is designed to support cost-effective and more power-efficient single-mode fiber interfaces for 100G and 400G Ethernet using 100G optical technology. A couple of other advancements include the 100G operation over dense wavelength division multiplexing (DWDM) systems. This standard is significant as being the first Ethernet specification of coherent DWDM technology supporting 100G connectivity over lengths of at least . Also notable is the 400G operation over DWDM systems, which extends the Ethernet specification for coherent DWDM technology to 400G.
These facets of Ethernet’s evolution are particularly interesting for communication service providers (CSPs). For decades CSPs have been driven towards higher-speed Ethernet due to their growing need for multi-service aggregation, which continues to expand with support for a variety both fronthaul and backhaul applications, thereby pushing Ethernet requirements for higher rates and longer distances. The escalating demands in X-haul connectivity, particularly long-haul connectivity managing immense amounts of data spanning the cosmos, are compelling data centers and CSPs to find transport solutions that are more flexible, reliable, and cost-effective. The response to this challenge lies within coherent optics, harnessing the power of light to enable compact and efficient transceivers to transfer enormous amounts of data over high-bandwidth, long-distance fiber connections.
For the first time, pluggable coherent optics for both outside and inside the data center are aligning, particularly with 400G Ethernet standards such as 400ZR. Designed to work in routers without constraining front panel space, these pluggable optics like Quad Small Form Pluggable Double Density (QSFP-DD) are quite efficient. Since 400ZR is an industry standard, the technology should theoretically be interoperable across different equipment vendors, certainly influencing and already altering how network operators design and construct their networks.
Figure 1- Transponder versus Digital Coherent Optics
As illustrated in Figure 1, thanks to 400ZR, IP-over-DWDM is now functional. This breakthrough allows for the placement of data center interconnect (DCI) optics directly into the router as digital coherent optics (DCOs). The removal of the transponder layer enables the management of optical layers to reside directly on switches/routers. At its most basic, coherent optics is a technology that employs low-cost standard optics to transmit 400G Ethernet over long distances using DWDM and higher modulation such as quadrature amplitude modulation (QAM). Potential use cases include hyperscalers’ distributed data centers, which can be achieved by simply installing a switch or router with coherent optics. This approach not only minimizes the physical space required for equipment and reduces power consumption but also could encourage numerous CSPs to rethink their network expansion strategies.
As we venture into higher speeds, the reality is that transmission over fiber becomes increasingly challenging, making coherent optics a preferable option. With digital coherent optics (DCOs), you'll attain improved economics and sustainability, alongside a fresh and innovative network architecture based on routed optical network design. This technology is compatible with standard switches and routers, and DWDM systems.
Figure 2 - Advantages of Digital Coherent Optics
However, DCOs will require the management of more data compared to traditional optics due to the software and intelligence residing on the optics modules. Therefore, the end goal is to automate as much configuration as possible, and the system must be conducive for machine-to-machine (M2M) communication. Additionally, the management and automation of DCOs should align with the industry’s push towards open systems and software defined network (SDN) architectures.
The Optical Internetworking Forum (OIF) is currently working on the 800ZR project, and it is anticipated to provide a long-haul 400G option. Implementing coherent optics within a data center may not be cost-effective at present, but it could become a viable option once we reach 1.6T (2x800G) capacity.
Ethernet, as a ubiquitous network technology, powers infrastructure all across the cosmos, including the new era of cloud-native 5G data centers that provide the infrastructure for 5G applications. Undoubtedly, 5G won't be the sole network technology capable of fulfilling all requirements. Therefore, wired and wireless networks will complement each other, with the wired networks requiring higher speeds to transport the wide array of new applications and services. We can expect to see Ethernet, this 50-year-old technology, reinvent itself once more.
The demand for higher speeds will also shape the evolution of Ethernet from a broader perspective, particularly in relation to Information Technology (IT) and Operation Technology (OT) networks. There remain gaps in how these “siloed” networks coexist and cooperate. Historically, these IT and OT networks haven't been well integrated. However, with Ethernet’s evolution, there is an opportunity to leverage advantages such as automation and other common tools. The ultimate objective here is to establish a single-protocol (Ethernet) network that caters to the needs of a broader range of infrastructures.