Reaching for 40GBASE-T

When developing a new networking standard, several attributes need to be balanced to optimize its implementation.  To optimize the implementation of 40GBASE-T, the task force developing the standard (IEEE P802.3bq) appears to have settled on a reach of 30 meters.  This is a tradeoff between power dissipation of the silicon physical layer (PHY) IC driving the cable, the complexity of the PHY which would impact cost, the implementation of the channel, and the reach of the link.

The question is: Is 30 meters long enough?  Let’s take a look.

Copper cabling, of one type or another, is the preferred media type for connecting servers to the first layer of switches.  Deploying a 10G Ethernet copper link with CAT6A cable and RJ45 connectivity, or using SFP+ Direct Attach Copper (DAC) cable assemblies, is less expensive than deploying those links with fiber optics.  One other problem with DAC cables is compatibility issues.  DAC cables have a memory chip on them that provides information such as the vendor – and this information is then sometimes used to shut down the port.  Structured cabling does not have this problem.

Putting aside 40G breakout cables (We will discuss this in a future blog post), the current copper solutions for 40G Ethernet limits the architectures one can deploy.  40G Ethernet QSFP+ DAC cable assemblies have a reach of 5 to 7 meters.  This means that one could use them to deploy  a Top of Rack (ToR) architecture or where one uses a ToR switch to service 2 racks of servers, but that is about it.  QSFP+ DAC cable assemblies cannot be used for most End of Row (EoR) implementations or other architectures.  Additionally, DAC cable assemblies are a point-to-point solution and therefore cannot support a structured cabling design.  A structured cabling architecture uses a permanent link, or backbone, and patch panels and patch cords that connect the permanent link to equipment.  This provides the maximum flexibility when it comes to deploying equipment and allowing changes in the future.

40GBASE-T provides a flexible, structured cabling solution for 40G Ethernet.

With a reach of 30 meters, 40GBASE-T is a perfect solution for deploying 40G in an EoR architecture, not to mention a ToR architecture.  30 meters is certainly long enough to implement virtually all implementations of an EoR.  Additionally, 40GBASE-T is a structured cabling solution.

EoR Implementaiton

The deployment shown in the figure uses a patch panel at the top of each rack and also at the top of the LAN cabinet.  Although the figure is depicting a 40GBASE-T deployment, it looks exactly like a diagram for a 10GBASE-T deployment.

There are numerous advantages when using 40GBASE-T over QSFP+ DAC cable assemblies:

  • 40GBASE-T can span the reaches needed to implement an EoR architecture
  • It is less expensive to deploy vs. DAC cable assemblies
  • No compatibility issues as there are with DAC cable assemblies
  • 40GBASE-T is a structured cabling solution
  • It is backwards compatible with 10GBASE-T allowing a pay-as-you-go CAPEX model

Our expectation is that the 40GBASE-T standard will be ratified sometime in early 2016.  If 40GBASE-T follows the same timeline as 10GBASE-T, then 40GBASE-T silicon, NICs, and switch ports should not be far behind.

We invite you to download our technology brief, 40GBASE-T Future Proofing Risks for more information about 40GBASE-T and the cabling solution that will support it.  Check back often as we will be updating the technical brief with new information about 40GBASE-T as it unfolds.

Leave a Reply

Your email address will not be published. Required fields are marked *