XKL on Middle Mile Networking Challenges

XKL’s Chad Lamb, Chief Systems Architect, and Darren Yuet, Marketing Manager, recently joined Sean Buckley on the Lightwave Broadband Pulse podcast to explore some of the biggest challenges in optical networking.

Introduction: Hi, and welcome to the Broadband Pulse, a new podcast series led by Sean Buckley from Broadband Technology, important LightWave. In this series, we will interview key executives from the broadband and optical industry segments about new and emerging trends in each industry segment. And now onto the episode.

Sean Buckley: Hello, welcome to Light Wave and BTR Broadband Paul’s podcast. I’m your host, Sean Buckley, Editor in Chief of Light Wave, and BTR. Today we have two optical networking experts. We’re gonna talk about some really great innovations at XKL. Joining me is Chad Lamb, Chief Systems Architect at XKL. Along with him is Darren Yuet, Marketing Manager for XKL.

Welcome, Chad and Darren to the program.

Chad Lamb: Thanks for having us, Sean. Appreciate it.

Sean Buckley: Yeah. All right. So here we go. With more interest in disaggregated approaches to network hardware current trends and management, what should businesses prioritize when evaluating their options?

Chad Lamb: Yeah, Sean. The disaggregated approach is really popular these days. It allows you to really optimize your purchases and make your decisions in terms of getting the best solution for that aspect of your network. It allows you to separate out and compartmentalize different aspects of your network. And so a disaggregated approach, as you said, is really popular these days.

The problem that is happening though is disaggregated management comes along with that. Software and the management of this hardware is oftentimes an afterthought. People are very much focused on design investigations. Designing a good, robust network, looking at features and capabilities — whether it’s Layer 1, Layer 2, Layer 3 — and the management aspect of it is left to the end. It’s not as important in the initial investigations and and testing.

Many people get caught up in the feature set at Layer 1. For network engineers, that’s always their biggest challenge, and they tend to focus on bandwidth and data rate and what are the flex-grid capabilities and ad-drop, and all these great features.

But the ability to manage and control that, it becomes a problem once they realize where they’re at. So their management philosophy that they currently have employed with their existing equipment, the switches or routers, servers, et cetera oftentimes does not work well when they try to integrate the optical layer equipment.

So when you’re choosing your vendors, it’s important to consider the entire ecosystem, the entire environment, not just the hardware and your hardware requirements and distances and reach and growth and capacity. But also — are you gonna be able to manage this equipment? How are you gonna be able to control your network and control your own destiny?

We hear from customers that they really are at the mercy of their vendors in some respects because they don’t know what to do in certain situations. They have to call up, get support, have somebody come out and help them with the problem. And network engineers really don’t want to have to rely on outside capabilities and skillset in order to to manage their stuff. And this can be a problem, not just with lower-end equipment. Maybe you’re buying something for very special use case and it’s a small piece of your network, it’s not an integral part of things from your network standpoint.

But these things really proliferate all the way up into the upper echelons of your network as well. Very high end transport gear, for example. The management aspect of this is often siloed. And one example of that, if you consider like an IP over DWDM solution, where you have network engineers taking the pluggables and putting them right in their router, they’re then forced or have to consider how are they gonna provide the amplification, the filtering, the flex-grid filters and these other components that go along with building up the rest of their network.

Many of these things come with a web interface for management or some JavaScript things that just don’t work well with their existing management environment and their management philosophy.

Sean Buckley: Darren, you wanna add anything?

Darren Yuet: Yeah, absolutely. Yeah. I think one more aspect of that when procuring equipment is just to do a internal audit when introducing new software, just to make sure it’s scalable across your network, but doing that within a controlled environment at first.

That means revisiting your conditional automations just to make sure those aren’t affected. Another thing to consider is user and role management as well, especially when certain permissions and access are usually given to specific users that have moved on to different parts of the org or are out of the organization.

So having good ownership of coverage even at layer one when changes aren’t as frequent. It’s always good to be able to stress-test your team and your equipment to ensure that the environment is healthy. One way to get ahead of this is to proactively ask for a mid file, for example, from your vendor just to see if their software’s based in SMP, and you can test compatibility ahead of time.

Sean Buckley: Chad, can I come back to something? You talked a little bit about IP over DWDM. Where is the market with that? I’m actually gonna talk to a carrier tomorrow about this topic, but how is that evolving from the XKL perspective? You mentioned some interesting challenges I know that people are grappling with, but there’s also benefits too.

Chad Lamb: Yeah, it is a double-edged sword. The pluggables, there’s a lot more common specifications at the higher data rates now when you get 400G ZR specification they did a pretty good job of defining what is needed to support that that kind of capability, so you have multi-vendor solutions out there so you’re not locked into a single transceiver vendor, for example. And 800G is similarly positioned where, for the baseline ZR interfaces, you’ll be able to support a common interface that will allow multi-vendor interoperability.

And so if your switch and router can support the footprint — the power consumption, tuning, a lot of the telemetry that these devices, these various complex devices, that have an awful lot of data like prefect post-effect counters, a lot of different things that you wanna monitor and manage, and if your switch and router can step up and handle all of these aspects of the pluggables — then the next step is okay, so now what do you do to combine all the channels that you’re gonna have with this and amplify? How are you gonna deal with, do you want flex grid? You can buy that. Basically a WSS-type mux/demux.

So they’re used to dealing with their switch and their router, and they’re managing now. They can get some optical stuff out of  their switch and their router, they can start using these pluggables. But then they go to a whole different management environment, to look at, monitor, and manage the amplifiers and the tunable filters or whatever the aspects of the DWDM network that they have to build on their own.

So IP over DWDM has its use cases but it doesn’t scale well. So if your channel count is modest and your data rates are modest you can really take advantage of not using a muxponder or transponder. And as long as you can manage the rest of it, it’s a reasonable approach that people can go with. It’s got a lot of dependencies. Like I said, if you get a 400 pluggable, does your router allow you to tune that? Does your router give you all of the telemetry that this device is providing? And so that’s another key aspect of this and you gotta be aware of it.

Sean Buckley: Interesting, that’s for sure. I’m gonna come back to the disaggregated approaches to the network hardware. How do those apply to network operators who have outsourced their management in terms of software and on the disaggregated side?

Darren Yuet: At least in my conversations with network engineers, in an ideal world, they still wanna manage their own equipment in-house. They don’t want to have to call their equipment provider to get support even on minor issues. I guess the benefit of outsourcing your network management is that it mitigates risk and you don’t have to dive deep into the GUI or be aware of any software that you’re not familiar with, your team’s not familiar with. So that comes at the cost of paying for that outside help through these support contracts. When a problem does arrive you still have to be in contact with that third party team to diagnose and resolve an issue and sometimes the timing is never right with those, so they become unpredictable.

But in our experience when support isn’t outsourced, our own customers have mentioned they really only interact with Layer 1 equipment at configuration and then during regular maintenance windows. That means that you can go months or years without using those GUIs provided by vendors, so if a GUI is unintuitive or overly complex, it becomes really time consuming and frustrating for a team that’s short staffed especially if those changes still require going into the CLI or the NMS if it’s not found in the GUI. That feedback definitely influenced our decision to offer a CLI versus a GUI.

So the interface looks and feels like a network box, like a router or a switch. So that engineers can make changes on the fly and it integrates right up into their NMS like Cacti, OpenNMS, IOS all of that managed through SNMP.

Sean Buckley: Keeping things smooth.  Okay, great.

Chad Lamb: Yeah and the other side of that, Sean, I ust wanna interject here for a second. Network Engineers are oftentimes a little afraid of this spooky Layer 1 stuff. They don’t necessarily understand the optics, they don’t understand how the light goes down the fiber, they know some of the basics. And managing these components that build your optical layer and your DWM transport can be a daunting task to them. And so that’s why it’s really important, as Darren mentioned, the CLI, for example, is very router-like. That’s great, right? You always have that foundation to work at. That doesn’t mean you’re gonna be there very often, but knowing that when that emergency arises, when all else fails, you need to be able to get down to the bare bones of the box and do something.

And so having that familiar CLI underneath all of this really gives the Network Engineer the confidence he needs to say, ‘Hey, I can do this. I can manage this gear. It’s not voodoo. It’s not magic. And I don’t need to be an optical physicist to be able to solve my problems.’

Sean Buckley: Yeah. Making it easy and seamless, sure. What are some of the complexities that service providers underestimate when designing a network? And I think a lot of our discussion is leading into that here, but do you both wanna take that on?

Chad Lamb: Yeah, that’s a great question. Outages obviously are the top of the list, right? Networks are designed to handle problems. They’re designed from the beginning to to work around issues. Things happen. Hardware fails. Backhoes take out fiber, right? Being able to plan for that is really important. And part of our approach to DWDM and providing solutions to customers is to consult with them and discuss with them these things.

They don’t necessarily know all of the components make up or can be in their optical network. They’ve heard of an OTDR, their fiber provider will give them an OTDR result when they lease the fiber, but they don’t necessarily understand that they can get that same feature in the integrated XKL solution, where if you do have that fiber break, the OTDR is just gonna run automatically and instantly deliver a trap. It’ll send you a text message, whatever you need, whatever is your operating mode for management. And it will give them that report instantly, and then they can start diagnosing and understanding, right? They can send their team, dispatch them out to that location, address the problem, get things up and running again.

Same thing with optical protection switching. This is a pretty common function at the optical layer, and oftentimes it’s approached as a disaggregated solution. When you aggregate it, when you integrate it all in, and you do that under the same management plane, that allows for the Network Engineer to easily take advantage of that feature, understand what it does, and and be aware and prepared for when that fiber does get cut. And the OPS switches over, uses the backup path, and he gets all the telemetry information he needs to deploy his team, to figure out what needs to happen next.

So it’s not just these great hardware features that are important to help providers have a robust network and make sure their customers don’t go down. They need to empower themselves to be able to manage this equipment too. So this part, the network design part, which we really help them out with and we give them this guidance like, ‘Hey, you got this OTDR feature, you got this OS feature integrated, this is how works. This is how we mitigate some of your concerns about the failure modes.’

Some customers, they want have redundant equipment all the way up. Pairs of switches, pairs of routers, pairs of Layer 1 stuff, pairs of fibers, pairs of everything. That’s one end of the scale. And you have some people on the other end of the scale, they don’t even have a pair of fibers. They have a single strand of fiber. So you really need to work with your customer closely to let them know what are their options and how can they design a robust network that they can manage and control.

Sean Buckley: Because a smaller telco versus let’s say a bigger customer, a big telco or a large cable company. Or a small, independent, I don’t know, let’s take an electric CO-OP getting into fiber to the home, or whatever it may be. All of those are different kinds of networks and they all have different topologies and different needs. I guess trying to understand the different nuances is what you guys are talking about, giving them that visibility.

How is XKL working with your service provider customers to solve these problems that both of you have laid out? Let’s talk a little bit about that.

Darren Yuet: Yeah, absolutely. As Chad was alluding to earlier we see a wide range of requirements and resources that companies dedicate to their networks, larger companies, as you said, Sean, like they have dedicated Optical Network Engineers, most likely who they can rely on as subject matter experts for their network architectures.

Some Tier 2, Tier 3 service providers simply don’t have the human resources to learn DWDM to the extent that they need to to make these critical decisions, so XKL can help them on the design side and even if their network is already entirely built-out, XKL helps works through blockers and provide options. Because teams aren’t really ever done building their networks or making enhancements, so the key from there is just avoiding extraneous or overkill solutions and understanding alternatives that they didn’t know existed at the time.

So an example of this is our work with Scatter Creek. They’re a local rural ISP here in Washington State. They were evaluating different solutions, but they had a specific architecture in mind already so their network team wanted to manage it in-house 100%. And part of their philosophy was being fully responsible for it, being able to understand what the software does and solve problems internally. So we matched them with a solution that was simple enough to understand, but also configurable to his liking because he likes to tinker with his network and be able to set those traps and alerts as we mentioned earlier. And even in our onboarding, in our training with them they took that in stride with virtual training, online demos.

And by the time deployment came around it was crickets, essentially. Our team was ramping up and waiting for their support calls, but they just never came. It was that easy, and they loved it. So the minimal training was enough. And, some vendors might quote, a five figure deployment cost for an onsite configuration, but we want to empower the engineers who wanna do it themselves to be able to do that, and it’s extremely gratifying that we ended up saving them, I think, 40% on their operating expenditures, which they were able to pass down to other local Washingtonians. So very fulfilling to be able to help out our local communities.

Sean Buckley: That’s the very nature of the smaller independents. Trying to get things done and that’s why they’ve been pioneers in things like fiber-to-the-home and broadband, etc., so that’s great to see.

Chad Lamb: Just for a moment here, if you push into that 40% OPEX number, what really that comes down to is the engineers have empowered themselves here. We’ve empowered them to now control their network. They can manage it. They can understand what’s going on without having to call us up to come out and say, ‘Hey we got this thing happening. I don’t know what happened. What’s going on here?’

They can do it. So that OPEX is oftentimes overlooked because you’ve got this staff and they’re working and how do you quantify those costs? By providing them this ability to keep the management plane out of these management silos so that you have an entire network here that’s gotta work in concert.

The Layer 3 equipment, Layer 2, Layer 1 — your servers, your sand servers, everything has to work together. And they need to be able to manage this stuff from a central facility or whichever way they like or prefer to manage it. And so that’s where that OPEX comes from, changes at the optical layer, the DWDM layer, as we said, they’re so infrequent that they happen maybe every couple months. Maybe every couple years, really. And so if you’re a network operator and you deployed this stuff and a couple years go by and you need to make an adjustment, or you want to do a firmware update, or there’s something you wanna, do you’re reluctant to get in there because you’re not sure you remember what to do.

You do something infrequently, right? You forget. And but by making the management interface familiar — they’re dealing with their switch, their router and their servers on a daily basis  — if we make things look like that equipment, but provide these robust DWDM capabilities, they can get in there every other year or whenever they need to, they’re not concerned about it. It’s very straightforward. They know that they’ll be able to pick it up really quickly. So I think that’s another aspect of the OPEX that’s overlooked oftentimes.

When you talk about this Layer 1 stuff, it’s expensive. These lasers are expensive, all this stuff. But what about the OPEX? That’s expensive too, when you really add it up.

Sean Buckley: I look forward to maybe following up with that customer at some point, but great. Thanks both of you. I’d like to bring this to a conclusion, do a little wrap-up with some final thoughts and how you see things shaping up for 2025. I hear there’s some good light coming out with some of this industry glut thawing out, but I’d love to just in general hear some final thoughts from both of you and see what we should expect next year.

Chad Lamb: Well certainly at the optical layer, 800G is coming. That’s a big lift for your router ports, but that’s not expected to be in many customer networks in 2025. But the DWDM side of that will be available in early 2025. And you don’t necessarily need to have an 800G port on your router to take advantage. You can do a lot of aggregation of your 100G. We can take 8 x 100G and feed it down an 800G on the lineside.

So you get some spectral efficiencies, you get to take back some of that spectrum that you had to burn at the lower data rates. And so even though the 800G router ports are not gonna be ready in 2025 — or I shouldn’t say they’re not gonna be ready, they’re out there. If you can stomach the cost, go for it — but 400G is still almost unobtainable for a lot of customers in terms of their router ports. So we’ve gotta get through 2025 with increased number of 400G router ports. Meanwhile, the 100G is really proliferating throughout the networks, so every layer, lots and lots of 100G. So at our layer, the optical layer, we’re happy to transport that 100G, but we can do it more cost-effectively and go further aggregating that up to 400G and then soon 800G. So that’s what I see happening in 2025 in terms of pluggables and support for higher data rate transmission.

Sean Buckley: And Darren, over to you and we’ll round out here.

Darren Yuet: Absolutely. Just to piggyback off of what Chad was saying, we’re not only looking at the bleeding edge here, but also finding opportunities for existing networks or networks that aren’t seeing extreme or bigger leaps in bandwidth requirements and helping them, those types of customers, maximize their existing fiber utilization as well. We’re still developing products for 10G. We’re still looking at other opportunities to use DWDM in cross-connects and data centers, for example. There’s a lot of opportunity, I think, well within the existing technologies, but just continuing to maximize our team and their expertise to help customers find ways to enhance their existing networks as well. Not just the bleeding edge application.

Sean Buckley: For sure. And that reminds me of some other conversations we’ve had recently too. Great, I thank both of you and I look forward to talking to you more as we wrap up the end of 2024 and going into 2025 and spending more time with XKL and your customers. It’s certainly a lot to talk about, but again, thanks both Darren and Chad for joining us today and happy holidays to everybody out there.

Looking forward to seeing how things unfold as we move forward into 2025, I think it’s gonna be an exciting year. Thanks again for your time and have a great afternoon. Thank you. 

# # #

About XKL

XKL helps businesses of all sizes expand and reimagine their networks in innovative ways with built-to-order, cost-effective and easy-to-manage DWDM solutions. Backed by our proven customer-centric, consultative approach and 30+ years of experience, XKL’s Layer 1 optical networking solutions deliver the capabilities and reliability required to meet even the most challenging network demands.