91桃色

On this episode of Living Live with 91桃色, host Tyler Kern was joined by 91桃色 Marketing Product Manager, Connectivity & Control Todd Riggs to discuss the ins and outs of decentralized IP and SDI routing.

While these aren鈥檛 new topics, a focus on IP transport topologies and a move to UHD and 4K productions have spurred a new look.

鈥淏oth of these things are stressing some equipment that we鈥檝e had, traditionally, so we鈥檙e having to work around some of those issues within the equipment,鈥� Riggs said. 鈥淏ut there鈥檚 also some promise with some of this that鈥檚 allowing us to explore different workflows.鈥�

In the past, the way to build a large-scale router involved leveraging combiners, tie lines and more that could facilitate those bigger infrastructures.

As technological advancements were made, these large-scale routers shrunk down to a single chassis. However, even these infrastructures didn鈥檛 have the bandwidth required for today鈥檚 applications.

Now, with decentralized IP, unprecedented distributed models are able to be scaled to sizes that weren鈥檛 possible in the past.
However, there is no one-size-fits-all solution, even for a frontier this promising.

鈥淚n fact, I would probably say that, for the majority of customers, a decentralized model probably doesn鈥檛 make sense,鈥� Riggs said. 鈥�(That鈥檚) simply because I don鈥檛 know people that have unlimited budgets, and cost is a factor. These models tend to be a bit more expensive up front just because of the way you build these out.鈥�

Still, these decentralized models are helping shift workflows by providing scalability and ways to link multiple sites together in a much more efficient manner than single-frame models.

AUTHOR: Tyler Kern
CONTRIBUTOR: Todd Riggs

Podcast Transcript

Tyler Kern:

Hello everyone. Welcome to Living Live! with 91桃色. I am your host, Tyler Kern, and today I’m joined by Todd Riggs. He is the Product Manager for Connectivity and Control at 91桃色. Todd, thank you so much for joining me today.

Todd Riggs:

Thanks for having me.

Tyler Kern:

Absolutely. So today we’re focusing on decentralized or distributed IP and SDI routing. Now this isn’t necessarily a new concept or a new topic, but Todd, explain to us why this is relevant for us to discuss now.

Todd Riggs:

I think it comes down to a couple of things. There’s a couple of really big shifts going on within our industry as a whole. They’re familiar topics to anybody that is working in our space. But they have changed some of the requirements and how we address some of these requirements within a routing topology. And those two topics are obviously IP transport topologies and then this move to a UHD or 4K type productions. Both of these things are stressing some equipment that we’ve had traditionally and so we’re having to work around some of those issues within the equipment. But there’s also some promise with some of this that’s allowing us to explore different workflows based on a few different factors. So it’s an interesting thing. It’s a bit of Back to the Future. A decentralized concept isn’t new. We were forced to do this years ago based on some limitations, and we’re coming back around to it now.

Tyler Kern:

Todd, you got into a little bit of it there, but I was wondering if you could go into a little bit more specifics and flesh out in a little bit more detail. The advancements in technology that have really caused everybody to revisit this particular architecture that, as you mentioned, is something that was done more in the past.

Todd Riggs:

Yeah. So if we look at the past, one of the big challenges that we had, and if you go back 20 years, the way you would build a large scale router, we couldn’t build these large systems into single frames. And so what that did is that made us build these larger systems through a series of things using things like combiners, things like tie lines, et cetera, to start to build these bigger infrastructures. About 15 years ago, 20 years ago or so, we started getting into bigger cross-point chips that we could now put onto cross-points within routing systems and now instead of literally taking an entire room to build out a 300, 400 ish sized router, you could now do into a single chassis and go up to a thousand or more. So it allowed us really to compress that footprint.

Todd Riggs:

Now what’s happening with this rise to 4K or 12G, UHD, whatever you want to call it, this has put a strain on that because most of those infrastructures were based around three gig topologies and to get into 12G or UHD production, which we’re seeing a lot of our customers want to do, a lot of times it’s going to require four times the bandwidth and those infrastructures couldn’t really support that. So one of the ways you can address that is through singling 12G equipment and Ross and other manufacturers are going very quickly towards moving a lot of their portfolios to that because that allows you to keep that same model.

Todd Riggs:

But the other side of the coin to this is as we are moving into this IP transport, there’s some interesting things we can do with that because there were some inherent limitations when you are using a co-ax for instance for distributed models. And where IP and really starts to get interesting is you can start to use the IP switch as a switch core, if you will, and you can start to build these distributed models that can support UHD and others and start to scale in a way that we were not able to do in the past. And so that’s why when you look at this combination of this decentralized model, the requirement for 4K, we can start to move and this topology is allowing us to move into some pretty interesting things that our customers can do.

Tyler Kern:

So I think it’s probably important to mention that these aren’t one size fits all solutions, right? That there are different factors that play into what makes the best solution and what makes the decision possible for each potential client. So walk me through those factors that you consider when you’re trying to decide what solution is best for each potential client or customer that you have.

Todd Riggs:

Oh, sure. So that is very true. In fact, I would probably say that the majority of customers, a decentralized model, certainly today probably doesn’t make sense simply because one of… I don’t know people that have unlimited budgets and cost is a factor. And so these models tend to be a bit more expensive upfront just because of the way you build these out. So you’re doubling your IO in a lot of senses and some other things to get this flexibility. So budget is obviously a big factor for these types of systems. But in terms of what is positive about these is it allows us to start to change the workflow a little bit. Instead of these big monolithic frames, we can start to move into a workflow that starts to link multiple sites together in a very efficient manner.

Todd Riggs:

So whether it’s multiple cities or multiple rooms or floors even in a building, you can start to use smaller building blocks and then start to link these together. And so that’s a really interesting thing for a lot of our customers as they are building campuses or they might be building multi-site facilities and they want to be able to tie all of these things together. And that was something that you really couldn’t do traditionally with a single frame that a distributed architecture allows you to do. So it doesn’t allow you to change your workflow. So if you are a customer that is looking for those types of workflows, it’s a really interesting thing to consider. The other thing is just simply scale. One of the biggest challenges that we have with routers, and we’ve had them since they’ve been around, is the cross-point does get limited.

Todd Riggs:

So you can build a system and it can be a 16 by 16 or it can be a thousand by a thousand, it doesn’t really matter in the size, but you’re always going to be limited to how you build out the cross-point matrix and up until you hit that limit, so a thousand, you’re fine. Once you get to input 1,001 or whatever that may be, you end up having to… That frame can’t support that, and so one of the things that a decentralized system does really nicely is scale. And so this is another conversation that we’re having with our customers. What are your needs now and then what are your needs three, four, five years from now as you look through these transitions? What are you trying to do? Because it may be that right now you might only need that single chassis, but in the future you might need to scale to a point where a single chassis starts to not make the most sense and you can start building those blocks now and it does allow you to scale pretty elegantly actually as you move forward.

Todd Riggs:

So you have to weigh all of those factors. You have to weigh what’s their workflow? You have to weigh where they want to go, and then you also have to weigh what their budget is and based on that, you can start to put together a decent solution that really benefits them the most.

Tyler Kern:

So Todd, I think the assumption is that the transport interlinks our IP, but that doesn’t necessarily have to be the case, right? They can be SDI or fiber. So walk me through those choices and explain these various options and what makes them viable depending on the circumstance.

Todd Riggs:

Sure. So a lot of that comes down to the application and how you want to build these out and there’s really two ways to build one of these, and I’ll use the term blocking and non-blocking. A blocking architecture means that you can get signals from one frame to another, but it’s not unlimited. So let’s say you’ve got two frames and they have 64 by 64 for instance, if you put 16 links between those two frames, what that means is that if I want to go take something from frame A and move it to frame B, I’ve got 16 links to move that across.

Todd Riggs:

Now for some people, they don’t need a lot of interaction between these frames and so you can do that in a lot of ways. You can do that through simple co-ax, which is probably the cheapest and simplest to implement. If it’s a decent distance, you might use fiber, especially if you’re doing UHD. Because of the cable reach, using co-ax isn’t as great as it has been with some of the lower data rates. Fiber makes a lot of sense where you can start to move this over really good distances. But it’s a blocking architecture and that I’m never going to be able to get all 64 inputs from frame A over frame B at the same time. But again, we have a lot of customers actually that that works just fine for them. They don’t really need this and they’re building facilities with five or six different rooms and they’re actually using fiber because they just need a bit here and a bit there and you throw all that together with a nice sophisticated tie line management system and it works great.

Todd Riggs:

When it comes to a non-blocking architecture, you can still use those topologies but it does get harder the more frames you add and without having a drawing, it gets a little bit difficult to explain. But the theory is the more frames you have, the less local IO in each frame you’re going to be able to use because you’re going to have to build some sort of a mesh network to be able to move those signals around. This is where IP really gets interesting for us and for some of our customers because in effect, what we can do is we can bring in a IP switch, basically almost to act as a cross-point and then you’re sending however many links from each of those frames to that core switch.

Todd Riggs:

And then you can, again, build this in a blocking architecture but you can build this in a completely non-blocking architecture. And what I mean by that is, if I use my example earlier, if I have 64 local inputs and outputs and I run 64 inputs and outputs or streams to the core switch from both of my frames, then I can get any source to any destination at any time. And so that’s what I mean by a non-blocking architecture and an IP core in the middle of this or an IP switch in the middle of this really allows you to start scaling. So if you need to add more inputs, you add another frame and you add more links to the core. If you want to add 10 more rooms in theory you can do that. And so that topology gets really interesting for these very big, very distributed systems.

Todd Riggs:

The downside of that is there is a cost associated with that right now as everybody knows. It is not the most cost effective to use IP right now in a large scale. It is expensive, but it does allow the most flexibility. And so we have customers actually doing both. We have some customers that have said, “You know what? Blocking is fine, and we can do this with fiber. Our staff knows this the best. It’s the same workflow. It’s very simple to set up. It works the way we expect it to.” And then we have others that are saying, “You know what? We’re covering great distances,” or, “We are starting to build out our IP plants and this makes a lot of sense for us,” and they have the support staff to manage that and make that work and everything else and they’re moving in that direction.

Todd Riggs:

So in this case, it’s a combination of what type of architecture are you trying to do? How flexible do you want it to be? And then in a lot of cases it’s are we ready to make the transition to IP? Because it does require some different skills that some of our customers have and some of them don’t. And they’re just trying to move there slowly.

Tyler Kern:

So Todd, it sounds like from just this conversation that we’ve had, that that communication between yourselves and the client is really important and really vital just in understanding what are they trying to achieve? What facility are they looking to have? What capabilities are they looking to possess once their facility is fully built out. And that, to me, seems like it’s a really crucial part of this whole thing is just understanding exactly what they’re looking for. So then you can find the solutions that work best for them.

Todd Riggs:

Yeah, that’s absolutely true. One of the great things that I’ve loved about being a Ross employee, we have a code of ethics and one of the very first ones is we’ll always do what’s in the best interest of our customer. And in this one, it’s really important because you really do have to spend some time with our customers to figure out what they’re trying to do. It makes no sense to put them into an architecture that they can’t support or is out of their budget or really doesn’t meet what they’re trying to do in the future. And so we do spend a lot of time up front, especially with these, because these are complex. Even if you’re simply doing fiber links between the two systems, they are complex.

Todd Riggs:

And so we do spend a lot of time with our customers going through the scope of the project, their workflows, what are they trying to achieve, the key things to make this work in their minds and, and obviously, what their budget is for these projects. So to make sure we give them the best fit. And if we have a fit for them, wonderful. If we don’t, then it’s okay to say we’re not the best solution. Hopefully we are. But you do have to go through that because there’s just too many things that just might not make sense for them.

Tyler Kern:

Todd Riggs, Product Manager for Connectivity and Control. Todd, thank you so much for joining me here on Living Live! With 91桃色.

Todd Riggs:

Thanks. Thanks for having me.