Why 100G Networking Is Suddenly A Day To Day Enterprise Problem

Evan Kirstel (0:01)

Hey everybody, really exciting and intriguing chat today with a real innovator in the networking world on why 100G networking is moving into the enterprise and how teams can actually test and troubleshoot it with Jeff and Thomas from NetAlly. Gentlemen, how are you?

Jeff O'Mara (0:19)

Super well. Evan, thank you for having us. We're excited to be here.

Evan Kirstel (0:23)

Great to have you here. Great to see you from sunny Seattle.

Jeff, maybe start with introductions to yourself and how do you describe NetAlly these days?

Jeff O'Mara (0:34)

Yeah, so I'm the CEO of NetAlly and we are a leader in the networking test and measurement space. And while tech is about the future, I want to just take a second to go back to the past. And we have a super duper proud heritage starting with Fluke and then through NetScout and then to NetAlly today.

And actually, this team is responsible, and Thomas in particular, for inventing the first hand-held network tester and also inventing the newest and most sophisticated hand-held network tester, which we launched last week called the LinkRunner 100G. And so we want to talk about that as well. But anyway, we're just super excited to be here and we're doing more and more in this space as AI is driving a networking super cycle. We're here to support the people that make those networks work.

Evan Kirstel (1:31)

Wow, what a storied history. And we're going to dive into all that. Really honored to have you both here. Let's start, Jeff. Let's dive in with you first. Why 100G? Why now?

Why is it even an enterprise issue? You've asked me five years ago, I would have laughed perhaps at you, but what's changed?

Jeff O'Mara (1:47)

Yeah, well, actually, what's changed, starting in 23, there were more 100G ports in the enterprise than 10G ports. And that's been growing and growing ever since. And so what hasn't really changed is the need for bandwidth grows and grows and grows. And I think you might remember back in 2000, oh, all these dark fibers, and are we going to be able to light them up?

The hunger for bandwidth and the need for bandwidth is inexhaustible. The more bandwidth there is, the more that's going to get used. And now with AI coming in, right, you see more compute, more data centers, that means greater demands on networks. And so it's not just the carriers that have these, you know, large bandwidth networks, they're in the enterprise. And if you're responsible for running those in the enterprise, when one of the bigger links go down, you're losing more traffic.

And so that's happening. And what hasn't happened is the different standards that make some of that work easier on 10G and below haven't been able to keep up with the technology. And so what we've done at NetAlly is actually do that innovation to make it really easy for network pros to work on 25G and up.

Evan Kirstel (3:10)

Amazing. And Thomas, being the expert here on the call, what gets harder when you move from 10G to 25G, 40G to 100G? How does the problem change? Sure. It changes because a lot of the technologies that have been in place for 30 years on Ethernet that have enabled it to be used fairly easily, aren't happening, aren't available at 100G. There's auto-negotiation of parameters and a link up. And it turns out that this 100G on Fiber, which is all on Fiber, takes you back to everything having to be hard-coded in the system on both ends.

And there are literally like 30 different ways a switch port might link. And one of them is going to work, or maybe two, and you've got to get it right. And there's a lot of parameters. And so you take that and then you add the complex physics that are going on with these 100G Fiber networks. They're very futsy. The little bit of grime that you might wipe off on your shirt on a 10G Fiber, you're not going to do that on 100G Fiber and have it still work. So from the physics to the complexity of configuration, those are the two big things.

Interesting. So what problems do you see most with fiber optics transceivers at these kinds of speeds? Where is the troubleshooting? How has it changed?

It's changed in that getting the configuration right is complicated and fraught. And for example, when you're configuring a switch, you're a network engineer, your job is to get the switch configured right, it used to be easier to get it right. Now it's hard to get it right. And the interfaces, the command line interfaces, didn't actually keep up with the complexities of 100G. There are vestiges of automation in the switch configuration interface, the command line interface, which makes no sense. You can set speed to auto on 100G port, but there is no auto. And so, it's really easy to get wrong. And if you carry forward your experience from 1G and 10G, you're like, I know how to do this. Well, a lot more care is required. So major problem on misconfiguration of the two endpoints. And so, that's number one. Number two is the fuzziness in the fiber world. And some of the tools related to getting the fiber right and knowing that it's clean and capable don't give you an answer anymore the way they have had for 1G and 10G fibers. So there's been some breakdown there. Our approach has been, at the end of the day, the network engineer needs to know, how do I link up to this network and get the data that I need? And so that's the problem, Luis. It was like, you know, forget these other details and we will tell you what it takes to make it work.