Extending 10 Gigabit Ethernet to the Edge

By Bill Woodruff

April 25, 2008

Today’s Gigabit Ethernet is fully exploited in the datacenter. Since Gigabit Ethernet (GE) began appearing in datacenters a decade ago, performance demands and computing strategies have changed dramatically. Server performance has increased 100-fold through higher clock speeds, more efficient and compact processor designs, and the use of multicore processors. Virtualization has become a key strategy for server management, demanding higher server I/O performance to support multiple Virtual Machines (VMs) per server. In addition, the concept of converged I/O has moved storage onto existing Ethernet pipes, further increasing the demand for bandwidth. These demands began to fill existing GE links years ago, and the unmet demand for bandwidth has begun to impact datacenter performance.

The obvious answer is to extend 10-Gigabit connectivity at the edge, between servers and their corresponding switch downlinks. However, 10-Gigabit connectivity for servers is materializing more slowly than desired, because optical 10-Gigabit Ethernet adapters are too expensive and too complex for use in the high volume environment needed for servers. The expectations and requirements at the edge of the network are not the same as they are for inter-switch links.

Short-term solutions exist for the bandwidth squeeze at the edge of the network. The first step to satisfying this bandwidth demand was link aggregation, where multiple GE links are set up between the server and switch, only incrementally increasing the overall available bandwidth. On the other hand, some users have sufficient bandwidth demand to deploy 10GE at the edge today utilizing XFP or SFP+ optical modules, and are willing to accept the cost and complexity inherent with such optical solutions.

10GBASE-T: High-Volume 10-Gigabit Ethernet

What the market is calling for is an implementation of 10-Gigabit Ethernet which is as simple and cost-effective as Gigabit Ethernet is today, with plug-and-play simplicity via familiar unshielded twisted pair (UTP) copper cabling and RJ45 connectors. Vendors of edge-connected datacenter equipment are eager to see dual-port NICs with servers connecting to aggregation switches, starting in the near future at the 16- to 24-port density.

The 10GBASE-T standard (802.3an, ratified by the IEEE in June 2006) provides this solution. 10GBASE-T leverages a couple of attributes inherent in silicon-based technology. Moore’s Law applies, delivering the benefit of continued process shrinks.

Integration is the other benefit delivered by Moore’s Law. The path to higher integration in servers will deliver LAN-on-Motherboard (LOM), in which 10GBASE-T appears as a standard feature, not as an option. Higher integration in switches will combine with dual-PHYs, which will be followed by quads and higher integration as process and circuit advances permit. On the way to this destination, 10GBASE-T PHYs will also be making progress on a number of vectors.

Power: The limitations for device power are derived from the requirements of the target platforms. PCIe Network Interface Cards (NICs) are limited to 25W, with pressure to see them come in at power of 18W or less. New generation 10GBASE-T (in conjunction with state-of-the-art controllers) meets this objective. Ethernet switches will aggregate this edge traffic.

First generation 10GBASE-T switches will be scaled to meet the demands of the datacenter. Densities of 16 to 24 ports are initially required, balancing switch fabric densities and desired oversubscription rates. These switch densities become feasible once 10GBASE-T achieves the low power required to fit within such a platform.

Over time, power will continue to be a key requirement. Servers will move 10GBASE-T on the motherboard, and switches will drive densities of 48 ports and higher. These improvements will be enabled as advances in semiconductor processes permit continued reductions in per-port power.

Costs: Copper PHY solutions provide the lowest total cost. 10GBASE-T leverages silicon rules, which will drive down costs. Perhaps more important is the nature of how copper PHYs are configured within equipment. It is difficult to conceptualize a server with anything other than a pair of RJ45s on the back panel. Similarly, Ethernet switches represent completely different cost profiles between copper and optics: line cards with RJ45s have no incremental costs as the capacity is utilized. Optics will not only require the selection and purchase of the correct module, but since the module is purchased retail, margin stacking will be present in the final customer cost.

Cabling Environment: Datacenters utilize “structured cabling,” which defines a cabling implementation where the racks and cabling are part of the “structure,” and the servers and switches can be placed (and moved) with great flexibility. Cross connects with patch cords provide the flexibility to connect any server to any switch. 10GBASE-T is defined for 100 meters, but as importantly, this includes 4 connectors, permitting a cross connect on both ends of the horizontal cabling. It is easy to visualize the benefit of using a cabling like UTP, which can be easily field-terminated.

Rate Flexibility: Customers expect connections over an RJ45 to always work. The standard defines auto-negotiation, which permits ports to link at the maximum possible speed, but most importantly, to always link. 10GBASE-T will coexist with 1000BASE-T (and 100BASE-T) for a number of years, and the triple speed capability of 100/1000/10GBASE-T PHYs translate to network operational savings because UTP links will always come up.

Green Initiatives: Low power may not be what comes to mind for most people when thinking about first generation 10GBASE-T devices. However, we will we see significant power savings for those with innovative architectures and through adoption of finer geometry processes. Further, there are a number of initiatives underway that will drive down system power. Most interesting among these is the work underway in IEEE802.3az, or Energy Efficient Ethernet (EEE). EEE promises to permit copper PHYs to see drastically reduced power under reduced traffic, but also to permit the benefits of lower power to extend within the server or system. Such new techniques are in addition to current green strategies that have been well proven with copper PHYs, such as wake-on-LAN (WOL). A 10GBASE-T link, which shifts down to the 100Mb/s speed, can enter a sleep state and use WOL to exit the sleep state via a network command.

A Marathon, Not a Sprint

New copper-based networking technology always follows a development curve that begins with making it work, advances to making it economically feasible for commercial deployment, and then drives on toward ever-greater deployment and ever-declining costs. We are now at the point with 10GBASE-T where technical feasibility is rapidly advancing toward commercially viable manufacturing and deployment. Semiconductor producers that win in the marketplace will be those that design their products to meet not only the key market and technical requirements for initial deployments, but to scale and improve as the market grows.

About the Author

Bill Woodruff is vice president of Marketing and Business Development at Aquantia Corp. (www.aquantia.com). Bill has identified and driven multiple generations of leading-edge interconnect technologies. He led Velio’s penetration of the grooming switch market. Earlier positions include GiGA (acquired by Intel), Vitesse and Mostek. He holds a BSEE from the University of Missouri and an MBA from SMU.

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion XL — were added to the benchmark suite as MLPerf continues Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire