NBase-ing Your Wireless Decisions

Posted on by

Cat5

Copper is heavy. I’m not talking about it’s atomic weight of 63 or the fact that bundles of it can sag ceiling joists. I’m talking about the fact that copper has inertia. It’s difficult to install and even more difficult to replace. Significant expense is incurred when people want to run new lines through a building. I never really understood how expensive a proposition that was until I went to work for a company that run copper lines.

Out of Mind, Out of Sight

According to a presentation that we saw at Tech Field Day Extra at Cisco Live Milan from Peter Jones at Cisco, Category 5e and 6 UTP cabling still has a significant install base in today’s organizations. That makes sense when you consider that 5e and 6 are the minimum for gigabit Ethernet. Once we hit the 1k mark with speeds, desktop bandwidth never really increased. Ten gigabit UTP Ethernet is never going to take off outside the data center. The current limitations of 10Gig over Cat 6 makes it impossible to use in a desktop connectivity situation. With a practical limit of around 50 meters, you practically have to be on top of the IDF closet to get the best speeds.

There’s another reason why desktop connectivity stalled at 1Gig. Very little data today gets transferred back and forth between desktops across the network. With the exception of some video editing or graphics work, most data is edited in place today. Rather than bringing all the data down to a desktop to make changes or edits, the data is kept in a cloud environment or on servers with ample fast storage space. The desktop computer is merely a portal to the environment instead of the massive editing workstation of the past. If you even still have a desktop at all.

The vast majority of users today don’t care how fast the wire coming out of the wall is. They care more about the speed of the wireless in the building. The shift to mobile computing – laptops, tablets, and even phones, has spurred people to spend as little time as possible anchored to a desk. Even those that want to use large monitors or docking stations with lots of peripherals prefer to connect via wireless to grab things and go to meetings or off-site jobs.

Ethernet has gone from a “must have” to an infrastructure service supporting wireless access points. Where one user in the past could have been comfortable with a single gigabit cable, that new cable is supporting tens of users via an access point. With sub-gigabit technologies like 802.11n and 802.11ac Wave 1, the need for faster connectivity is moot. Users will hit overhead caps in the protocol long before they bump into the theoretical limit for a single copper wire. But with 802.11ac Wave 2 quickly coming up on the horizon and even faster technologies being cooked up, the need for faster connectivity is no longer a pipe dream.

All Your NBase

Peter Jones is the chairman of the NBaseT Alliance. The purpose of this group is to decide on a standard for 2.5 gigabit Ethernet. Why such an odd number? Long story short: It has to do with splitting 10 gigabit PHY connections in fourths and delivering that along a single Cat 5e/6 wire. That means it can be used with existing cable plants. It means that we can deliver more power along the wire to an access point that can’t run on 802.3af power and needs 802.3at (or more). It means we don’t have to rip and replace cable plants today and incur double the costs for new technology.

NBaseT represents a good solution. By changing modulations and pumping Cat 5e and 6 to their limits, we can forestall a cable plant armageddon. IT departments don’t want to hear that more cables are needed. They’ve spent the past 5 years in a tug-of-war between people saying you need 3–4 drops per user and the faction claiming that wireless is going to change all that. The wireless faction won that argument, as this video from last year’s Aruba Airheads conference shows. The idea of totally wireless office building used to be a fantasy. Now it’s being done by a few and strongly considered by many more.

NBaseT isn’t a final solution. The driver for 2.5 Gig Ethernet isn’t going to survive the current generation of technology. Beyond 802.11ac, wireless will jump to 10 Gigabit speeds to support primary connectivity from bandwidth hungry users. Copper cabling will need to be updated to support this, as fiber can’t deliver power and is much too fragile to support some of the deployment scenarios that I’ve seen. NBaseT will get us to the exhaustion point of our current cable plants. When the successor to 802.11ac is finally ratified and enters general production, it will be time for IT departments to make the decision to rip out their old cable infrastructure and replace it with fewer wires designed to support wireless deployments, not wired users.

Tom’s Take

Peter’s talk at Tech Field Day Extra was enlightening. I’m not a fan of the proposed 25Gig Ethernet spec. I don’t see the need it’s addressing. I can see the need for 2.5Gig on the other hand. I just don’t see the future. If we can keep the cable plant going for just a couple more years, we can spend that money on better wireless coverage for our users until the next wave is ready to take us to 10Gig and beyond. Users know what 1Gig connectivity feels like, especially if they are forced down to 100Mbps or below. NBaseT gives them the ability to keep those fast speeds in 802.11ac Wave 2. Adopting this technology has benefits for the foreseeable future. At least until it’s time to move to the next best thing.

8 thoughts on “NBase-ing Your Wireless Decisions”

  1. Pingback: NBase-ingYourWirelessDecisions - Tech Field Day

  2. Tom, I disagree. This is technology for technology’s sake not a real world need. 802.11ac enterprise deployments won’t achieve peak speeds or even close to 1Gbps backhaul throughout 99% of the time.

    Here is my take:
    http://www.revolutionwifi.net/blog/2015/1/multi-gigabit-ap-backhaul-do-you-need-it

    And Marcus Burton had a great write-up as well:
    http://www.theruckusroom.net/2014/02/will-80211ac-stab-you-in-the-backhaul.html

    Cheers,
    Andrew

    Reply ↓

  4. Tom,

    Thanks for article (and for TFDx in Milan).

    A few comments I’d like to make – take them as you will 🙂
    1. Both the NBASE-T Alliance and IEEE 802.3 are targeting 2.5 Gb/s and 5Gb/s over CAT 5e/6, not just 2.5 Gb/s.
    2. IEEE 802.3 (http://grouper.ieee.org/groups/802/3/) is responsible for producing Ethernet standards. It’s the SDO (Standards Development Organization).
    3. The NBASE-T Alliance (http://www.nbaset.org/alliance/) has a specification for 2.5Gb/s and 5 Gb/s Ethernet. We will be building consensus for our spec, supporting pre-standard deployment, and advocating for our spec in IEEE 802.3 (as well as other places). The Alliance is a non-profit group of companies that are collaborating on a specification to solve what we see as an industry and user need.
    4. You said “Long story short: It has to do with splitting 10 gigabit PHY connections in fourths and delivering that along a single Cat 5e/6 wire.”. Probably my fault that I wasn’t clear. We will still be using all 4 pairs in the cable (like 10GBASE-T http://en.wikipedia.org/wiki/10_Gigabit_Ethernet#10GBASE-T), but slowing the symbol rate (amongst other changes) to ~¼ (for 2.5Gb/s) or ~½ (for 5Gb/s) compared to 10GBASE-T.
    5. You said “The driver for 2.5 Gig Ethernet isn’t going to survive the current generation of technology”. I agree, that’s why both the NBASE-T Alliance and the IEEE study group (http://grouper.ieee.org/groups/802/3/NGEBASET/index.html) are addressing 2.5Gb/s and 5Gb/s rates so we have a bridge to the day we really need 10Gb/s (maybe with 802.11ax?).
    6. You said “When the successor to 802.11ac is finally ratified and enters general production, it will be time for IT departments to make the decision…”. I disagree. We have data (e.g., this presentation I made to the 802.3 Study Group in January 15 http://grouper.ieee.org/groups/802/3/NGEBASET/public/jan15/jones_ngeabt_04c_0115.pdf) that shows that most of the installed base of cable today is cat5e/6. If you assume ~3 years refresh cycle for APs, ~6-7 years for switches and ~10-15 for cabling, the IT guys need to start moving now so they will be done in 5-10 years. Otherwise we will be having this discussion again in 10 years’ time. In the Panel session at Cisco Live (PNLCRS-1000 – https://www.ciscolive.com/online/connect/sessionDetail.ww?SESSION_ID=82252) , Jan Demey from UZ Leuven (http://en.wikipedia.org/wiki/UZ_Leuven) stated they have been moving towards Cat 6a for a number of years, but they still only have ~1/4 Cat 6a outlets. Except for new buildings and renovation projects, re-cabling a live facility is a big ask. You don’t want to be rushed into it.

    Thanks again. Hopefully these comments move the discussion forwards.

    Cheers
    Peter

    Reply ↓

    • Peter,
      Truly thanks for your reply! I appreciate the additional information. And to clarify my earlier comment (made from my phone, so pardon the previous brevity), is that I don’t see 802.11ac Wave-2 as the immediate driver for 2.5/5 Gbps as it is made out to be.

      I definitely see the long-term need and think developing this standard now makes absolute sense. I just caution customers who think they need it ASAP once they start buying Wave-2 wireless APs. I would advise them to upgrade switches when they have a natural lifecycle need or for other reasons, and at that point consider 2.5/5 Gbps to provide some semblance of planning for the future. But for the reasons listed in my blog post and Marcus’ blog post the performance of Wi-Fi (as MUCH as I love it) is almost never going to saturate a 1 Gbps AP backhaul link in an Enterprise environment or a mixed client environment. Maybe in the future it will once client capabilities change substantially, who knows.

      Right now I see NBase-T as a niche solution for APs in low-density areas that can utilize 80-160 MHz channels without causing co-channel interference AND also have very high capability clients (3SS laptops) AND have an application need for very high bandwidth requirements. That’s a lot of “if”s in my opinion and I rarely see it in enterprise networks.

      Thanks,
      Andrew

      Reply ↓

      • I should probably also add the caution of adopting a switching platform based on pre-standard technology. If the final standard isn’t compatible, they will be stuck with it for a long while!

        So in short, my view is:
        – No immediate need, 802.11ac Wave 2 is not a driver
        – Risk of obsolescence and incompatibility with final standard
        – Developing the technology and ratifying the standard now is wise
        – Customers should be planning future switch lifecycle with this in mind, but wait for maturation
        – For the rare instance where low-density AP deployments make 80/160 MHz channels feasible, there are no Wi-Fi neighbors, there are only a few high-end clients with demanding bandwidth apps, then this could make sense. But then again, in these types of situations re-cabling a few select APs is more realistic as well, so two cable drops likely more cost effective for a few APs than buying a new expensive switch.

        I’ll keep my eye on it; I’ll lab it up because I’m a techie after all; but I’m advising customers to hold off.

        Cheers,
        Andrew

        Reply ↓

  5. Pingback: Cisco NBase-T – lipstick on your Cat5e pig | @hansdeleenheer

  6. Pingback: Cisco Live Europe: NBASE-T progress report | Just another day at the office...

Leave a comment