This chapter features a deep dive into the transceivers specs that make 10G to 400G transmissions possible.
Would you like to learn more? Download our white paper on this topic here.
- Chapter 1 - Introduction to challenges for polarity in data centers
- Chapter 2 - Understanding MPO polarity
- Chapter 3 - Knowing the standards
- Chapter 4 - Optics from 10G to 400G
- Chapter 5 - Disadvantages of Method A TIA-568
- Chapter 6 - Analyzing Method B TIA-568
- Chapter 7 - Analyzing Method C TIA-568
- Chapter 8 - Introducing the multi-path solution
- Chapter 9 - Real-world examples of how multi-path works
- Chapter 10 - The advantages of a multi-path solution
Josh: Rick, this is great. I just want to make sure I'm following you because this is exactly how I came out of the show and this is a lot from talking to IT professionals and not necessarily presenting programs, but there is a lot of uncertainty on ... but there's certainty that the speeds are gonna get faster. That's clear.
Josh: Historically. So we know that. There's uncertainty in how it's going to be done.
Josh: And we've seen this, for example, the 24 fiber MPO ferrule.
Josh: You had to have it. Well, no, you didn't. Add into that some proprietary technology, namely Cisco Bidirectional Transceiver. Hey, it's a great piece of technology, but potentially a Band-Aid, but also outside of the standards, which is very important-
Josh: -or certainly a factor. Now add that into the market we know and we love that there is multiple different ways you can do this. Cost is huge. And, listen, you and I know better than anybody, as well as all of customers, adding cost in infrastructure every generation of hardware upgrade is a nuisance to say the least.
So I guess, really, that goes right into what I'm looking for you to present is what's the solution in this environment of uncertainty of disruption? We know things are going to go a certain way, but nobody really knows how we're going to get there.
Josh: Nobody's got a crystal ball. So how the heck do you do this?
Rick: Well, here's the thing. So, again, we talked about the different style of transceivers. Actually, if you guys want to take a look at your screen right now, I have obviously the transceivers that are on there. Obviously, you see the MPO-based optics. You see the bi-di-based optics, along with what maybe looked like duplex LC as far as bi-di goes and/or wave division multiplexing goes. There's obviously a push for 25 gig and 50 gig over duplex style signal. That's where it's headed.
But case in point, and you've got to think about it, right? So you have an existing data center because everybody's not building that new, right?
Rick: Okay. So you have multiple or multiple millions of dollars invested in your infrastructure utilizing a certain style of transmission technology right now. Say you're using bi-di, for example, and then now you've bought the latest, top-of-row switches and you've got the latest core switch and all the sudden, boom! 100 gig up links that are using SR4 or QSFP, MPO-based optics. How do you connect that now?
"Well, sure. I'll order an NPR cable. I'll plug it in. But not I've got to go out to, say, my top-of-rows that have LC-based style interconnects."
Well, now all the sudden, I'm switching over from a parallel optics system over to a duplex system and back and forth. Simply ordering a standard cable that we're accustomed to is not going to work anymore.
Say, for example, again, it's all duplex-based interconnect. Now I'm expanding out. I'm bringing in new pods, which is a big trend out there right now, building data centers out with pods. And I need to connect this into my system and this is all, again, 40 and 100 gigabit-based MPO optics. And I need it connected into my existing infrastructure, even if I have an MPO-based infrastructure with consent modules, you would think that'd I'd be able to, just for example, pull those consent modules out and put MPO features in and connect that moving forward. And that's simply not the case.
Number one, it depends on the method that you're using. Number two, you have to worry about the genders. And then, number three, I can't order the same MPO cable to plug in from, say, machine to machine or in structured cabling because I have to be concerned with the genders and the path. It becomes very difficult to do, very, very difficult to do.
And then, again, the benefit to some of the LC-based, higher speeds, whether we're talking about 25 or 50 gig, or if we're talking about what's going to be new over shortwave division multiplexing, is that the advantage of trying to stick with RLC-based infrastructure is that obviously we can run it on a duplex path or our duplex system. If you think about it, if we use MPO, we have to add lanes or fiber, for example, to our system.
Rick: Hence, more cost.
Rick: But, again, as you I think eloquently put, your infrastructure needs to be flexible because it's in our opinion and the majority of the industry's opinion that you're not going to be able just to settle on what technology. You're gonna have a mix of both duplex and parallel, and your infrastructure has to be flexible enough to go back and forth from parallel to duplex signal and you need to be able to do that cost-effectively.
And besides CABLExpress and the multi-path system that we'll go over in a bit, there's simply no way to do that on the marketplace right now without spending big dollars.