When speeds and feeds really matter – NBase-T and the Modern Office part 1

Welcome back to rsts11. With the conference season on pause for a bit, we’ll be catching up on some coverage from last fall. Look for fresh homelab posts, a couple of device reviews, and more. The who-I-work-for disclosure is at the end of the post. The second part of this topic is at Rolling Your Own NBase-T Network.

What is NBase-T and Why Do I Care?

Before I get into my story, let’s cover a couple of the basics.

NBase-T is a technology standard that allows faster-than-gigabit but not-necessarily-10-gigabit connectivity over Cat5e or Cat6 cabling. The NBase-T Alliance website says “close to 100%” of enterprises run Cat5e or Cat6 as their cabling plant. So with this technology, many to most enterprises can grow beyond Gigabit Ethernet at typical building cable run distances without upgrading to Cat6A.

The typical deployment of NBase-T today would have standard gigabit ports alongside multi-rate ports that should support 10/100/1000 as well as 2.5 and possibly 5 gigabit speeds. Some hardware may scale up to 10 Gigabit as well. This would allow aggregation of bandwidth beyond gigabit speeds without replacing cable or hacking hardware. But more on that soon.

Having spent the last 2 1/2 years in technology sales, I’ve heard thousands of people say that “speeds and feeds” don’t matter or aren’t important. If you’re trying to get across to a line of business manager or a C-Level who may not be as much a technophile as CTO/CIO types used to be, that may be true, but when you’re trying to solve a network physics problem, they do still matter. Let’s see how and why.

What is NBase-T and why did Robert care?

peter-gideon-robert-interop-2016

Peter Jones, Gideon Tam, and your author, in the keynotes at Interop Las Vegas 2016

I first met Peter Jones, chairman of the NBase-T Alliance, Cisco Principal Engineer, and Man With A Lot To Say(tm), at Interop Las Vegas back in 2014 (I think). At the time, I was still a first-person problem-solver (i.e. an IT practitioner who had to fix his own stuff), so as I talked to Peter and his alliance partners about NBase-T, I tried to fit it into my past networking problems and figure out what it meant to me.

At the time, I still had fresh memories of my efforts in 2007-2010 to grow a 50 person office in an old building in San Francisco to support 150 increasingly data-hungry employees. We had fairly modest (read: crappy) wiring with limited port density and some HP 4000M 10/100 switches that, before I got there, didn’t even have gigabit uplinks. After I built out the datacenter, I upgraded to gigabit switches with LACP links, but still only had a few dozen client ports to work with.

??????????????

One of my fondest memories of my first couple of Interop events, aside from airline chicken – the DS108 dual speed hubs in the media room. I still own some of these.

The thought of getting more gigabit ports to each pod or cave in the office sounded pretty good, and being able to drop a 5gig link to a little switch where I originally used a standard 5-port or 8-port Netgear or Linksys gigabit switch… well, that sounded really good. It didn’t do much for having 100+ individual gigabit ports squeezing out a 10 megabit Internet connection, but it could speed up media backups to our internal infrastructure. And I really wanted a 5gig card for my Precision Workstation that used to serve our trouble ticket system.

At that Interop event, a lot of the NBase-T gear was prototype or samples, nothing that you could go online and order from your hotel room that night. There was a drop switch that took a multi-gigabit uplink and fed a few gigabit links downstream to your clients, and of course some PCIe NBase-T NICs for servers or workstations. But nothing on the shelf to buy yet.

In the past two and a half years, in addition to my becoming a (somewhat distant down Tasman) coworker of Peter’s at Cisco in our respective day jobs, I’ve also watched the NBase-T technology grow and mature and get to the point where you can buy some of it without a secret handshake and two commas in your checkbook balance. It’s also become clear that my original reason for wanting NBase-T deployment was not what was driving the business. To Peter’s credit, he very generously and politely disavowed me of this idea early and often.

What should I want NBase-T for, then?

This might not seem obvious, especially if you’re not a network or wireless administrator, but the killer app for this sort of faster wired networking is faster wireless networking.

Think about that for a moment. The call is coming from INSIDE THE OFFICE.

With dual band 802.11n and proprietary extensions/devices, you could theoretically get gigabit wireless speeds, but it wasn’t until 802.11ac that you saw standards covering about 1.3Gbps raw speed and nearly a gigabit of usable speed. Wave 2 of 802.11ac added more channel width settings bumping the theoretical speed to about 3.5Gbps and the realistic (optimistic) speeds up to 1.5-2.3Gbps.

That sounds pretty nifty. Think how fast your cat videos will come down over a 20Mbps DSL or cablemodem link with a 2+ gigabit wireless link. But as silly as that sounds, that’s not the worst problem with the higher speed networking.

Look at the back of your AP. See that gigabit Ethernet port? There’s your problem.

No amount of wireless technology within your LAN will change the outbound speed, but with a lot more communication going on within the LAN (even a home LAN, with streaming both ways and backups and 30+ mobile devices for a family of 3), you may want to be able to feed your full wireless capacity throughout your LAN to enable on-premises business applications, fileserver access, collaboration, and high efficiency cat videos.

Going beyond that, enterprises using network optimization and/or caching technology can then make more efficient use of their bandwidth while giving better end user experience when your 30,000 employees all want to upgrade IOS within 5 minutes of the launch.

With NBase-T in your infrastructure, you can feed high performance workstations as well as higher performance wireless access points on a selective basis without running new cables or slumming it with 802.11n. You may still need a couple of new runs of Cat6A or fiber for 10 Gigabit uplinks, unless you put your NBase-T switches in the closets and fan out from there, but that’s definitely better than rip-and-replace for the whole building.

So where do we go from here?

In the next part of this conversation, we’ll take a look at some of the NBase-T gear I’ve gotten some exposure to in the last few months, how you might use it, and what a deployment might set you back if you’re looking to do this on your own credit card and a typical e-commerce website account.

If you have questions or comments on NBase-T, feel free to chime in on the comments below.

Disclosure: While this blog is independent and predates (by 3 1/2 years) my working for a hardware vendor, I do work for Cisco, who owns Meraki. Both Cisco and Meraki have NBase-T gear on the market. My coverage here is not influenced by that affiliation, and wired and wireless networking are not a part of my day job at Cisco.

3 thoughts on “When speeds and feeds really matter – NBase-T and the Modern Office part 1

  1. Nice post @gallifreyan 🙂

    One very minor nit, I think we first met when NBASE-T first went to Interop, and that was 2015 (The Alliance formed in late 2014). I guess this is like #DogYears, feels longer than it really was.

    Thanks
    Peter Jones
    NBASE-T Alliance Chair.

    Like

  2. Pingback: Looking ahead into 2018 with rsts11 | rsts11 – Robert Novak on system administration

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.