spirent.com

25 Questions to ask About 25G Ethernet Revisited

Man wearing glasses, with chin in hand on a question mark background

  1. 25G Ethernet? Do you mean 25G electrical lanes used for 100G Ethernet? Nope. This is 25G Ethernet. It has 25G chip-to-module electrical capability to thank for its existence, but this is a true 25G Ethernet "pipe". Think of it as 10G Ethernet sped up 2.5 times.
  2. Where did this come from? I don't see it in the latest IEEE 802.3 standard! 25G Ethernet was initially an effort by a handful of data center industry heavyweights to move the technology forward. It is now an IEEE standard 802.3by.
  3. There are two specifications? Do they interoperate?  The industry has a strong desire to provide interoperability to the greatest degree possible. Also, it's likely that most devices will support both specifications.
  4. Why 25G? Moving from 10G to 40G is a big jump, and it turns out the incremental cost of 25G silicon over 10G is not that great.
  5. I just moved my server interconnects to 10GE. Do I really need 25GE? Probably not for a while. Most of the early adopters will be Hyperscale Cloud providers and those companies that make their living in the cloud or that require extreme performance from their networks (e.g. high-frequency trading).
  6. Should I be looking for 25G interfaces on my next WAN router? Not really. Most of the early applications are in the data center.
  7. Does it support copper? Yes, but not BASE-T. There is a separate effort for 25GBASE-T. DACs up to 5m in length are part of the Consortium and 802.3by specifications.
  8. Does it support optical? Yes, the 802.3by specification includes an optical PMD.
  9. Does it support short reach and long reach optical? Yes, and no. The IEEE optical PMD is short reach (up to 100m). However, there are some vendors working on 2km reach optics over single mode fiber.
  10. Which interface form factors are used? QSFP28 is used for 4x25GE and SFP28 is used for a single 25GE port.
  11. Are these form factors related to QSFP+ and SFP+? Yes. The new form factors are the same size as their older counterparts but designed for 28G lane speeds vs. 10G for the older form factors.
  12. 28G? Is that a typo? Nope. Ethernet is not the only application for these form factors so there may be other higher-speed interface uses for them.
  13. Are these the only form factors in use? The industry standardized on these but there may be other solutions as switching fabric capacity increases beyond what QSFP28 can support in switch faceplate density. With 50G electrical lane interest ramping up for QSFP, we now see an interest in SFP56, QSFP56, and QSFP56-DD.
  14. 100GE SR4 and CR4 specifications require RS-FEC. Does 25GE need FEC? Yes and no. There are two FEC modes (BASE-R FEC and RS-FEC) and both are optional copper interconnects (DACs).
  15. Do I need FEC? What are the advantages and disadvantages? There is a latency penalty when using FEC. If you are designing for HFT applications this will be important to you. The benefit of FEC is it greatly reduces the number of uncorrected errors across the media and helps to extend the useable reach of those media.
  16. What kind of latency penalty are we talking about? It's estimated to be between 80ns and 250ns for BASE-R and RS-FEC respectively.
  17. Are there other disadvantages besides latency? There is a very small power consumption hit estimated between 25mW and hundreds of mW per port. Not a lot but it may matter to someone with over a million servers to connect!
  18. Can I use 10G or 40G DACs in my 25G network? No. 40G DACs are not designed to run at 25G per lane speeds (they run at 10G speeds).
  19. Can I split my 100GE interface into 4x25G? Many Top of Rack (ToR) switches are designed for operation at 100, 50, 40, 25 and 10G speeds. Some of these switches have a modular architecture and allow the user to install the desired blade based on speed and form factor. Various breakout cable options like QSFP28 to 4xSFP28 allow users the flexibility to customize their server to ToR topology.
  20. 50G? When did we start talking about 50G? Did I forget to mention the Consortium specification supports a 50G Ethernet interface made up of 2x25G lanes?
  21. So, is there an IEEE 50G Ethernet specification? Yes, the task force is called 802.3cd.  The objective is to define single lane 50Gbit/s.
  22. Will the IEEE 50G be a 2x25G design? No.The industry is trying to get to the next level of efficiency by defining a set of 50G serial standards. 802.3cd task force will address this requirement.
  23. When will all these standards be done? The IEEE has standardized 802.3by was completed in the first half of 2016. Any 50G serial standard is likely out in 2018.
    25 questions about 25G revisited blog diagram
  24. What kinds of 25G ports are on the market today? The first GA products are starting to ship right now. These are ToR switches, server NICs and various interconnects such as DACs (Direct Access Copper) and optical interfaces.
  25. Can we do 50 questions for 25G and 50G? Let's save the extra 25 questions for another time! But if you do have more questions please put them in the comments below.

To learn more please visit: https://www.spirent.com/Products/TestCenter/DataCenter.

Join Spirent’s BOOST program, your portal into the industry’s most advanced Ethernet performance and functional test solutions.

 

comments powered by Disqus
× Spirent.com uses cookies to enhance and streamline your experience. By continuing to browse our site, you are agreeing to the use of cookies.