(smooth music) >> Hi, welcome to this CUBE Conversation. I'm Lisa Martin. We've got James Wynia here with me, the Director of Product Management at Dell. We're going to be talking about modern data center networks. Jim, welcome to the program. >> Thank you, Lisa. Great to be here. >> So let's talk about this. We've had so many dynamics going on in the last 15, 16 months, I've lost count. A lot of dynamics in play that are contributing to IT complexity. There's new sources of data. We had this massive shift to work from home, work from anywhere, that's now kind of this hybrid environment. Talk to me about some of the core requirements of a modern network infrastructure that organizations need to deploy. >> Absolutely, and thanks for teeing that up. The modern networking requirements these days, so many people have moved home, and so as a result, then the infrastructure back on the farm, back in the data center have to be beefier. You have to have more capacity. You have to be able to handle more scaling operations. And so things like the ability to radically increase your backbone just by swapping in some different transceivers, possibly some different switches to support those faster transceivers, allow for us to multiply that bandwidth very quickly. So that's been a big result of what we have seen coming out of all this, the COVID madness. >> Yeah. (chuckles) Madness is a great description for it, and there's going to be that hybrid as we go forward. There's going to be that need to, for any industry, I imagine, to enable work from anywhere. But talk to me about where customers are from a speed perspective. 100 Gig, that's really mature at this point. Is that where most businesses are? And then what's the next step from there? >> Another great question. (chuckles) I mean, 100 Gig is pretty much the de facto standard at this point. It has really become very cost-competitive and very stable. I mean, we've really been shipping QSFP28 at the latest 100 Gig for five years, and it has become the de facto standard for many, many different scenarios. As we move forward, though of course, we just need to move more data is what it comes down to, and so the next logical jump from 100 is 400. And so 400 started rolling out about the time that COVID came on, a couple months before that. And so honestly, there was a slight kind of delay in the industry as COVID kind of made everybody take a step back and say, "Whoa, hold on." But now it's really come back in full force. >> So what does an organization, and we'll kind of just leave this as any industry, need to do to be able to prepare to go from 100 to 400, because as you mentioned, the data sources aren't diminishing. It's only going to continue to increase. >> Absolutely. And so one of the things is to make sure that the backbone infrastructure can handle 400 Gig. Ironically enough, the actual optical cable trunks, those are pretty much the same. And so if you were running single-mode fiber to go a long distance, you would use that same cable. So you don't have to rip out all your cable infrastructure. What you have to look at closely is when you plug that transceiver into a switch, what is it capable of running at? In olden days, that was probably 100 Gig. Now you have a 400 Gig, so you have to make sure that you have just the right hardware to go with that. And then as you go down the chain, down the stack, rather, from those, the switch from the cord, or the switch all the way to your server, on the servers we see a lot of interest in 100 Gig, even up to 200 Gig today. And so it's the same discussion. You're taking a close look at your NIC or your adapter. What is it rated at? Is it going to be able to handle a faster speed? >> So it's not a rip and replace. Can you give us an idea of the migration path that a customer would take, and how Dell would facilitate that? >> Absolutely. And so we have some great customers who have really stepped out in different ways. You have the Greenfield customers who, they're building out a whole additional data center, say. And so they would just, from the ground up, replace it with the latest and greatest equipment that's just already ready to go. Other customers that are just extending, maybe they're tapping a couple data centers together and replacing those 100 Gig links to aggregate them with 400 Gig links. And then they would maybe migrate, adding in an additional 400 Gig links down through the stack as it makes sense. So ethernet is ethernet, right? Whether you have 100 Gig on one link, 400 Gig on another link, it all plugs and plays nicely. And so you don't have to have this big step where you have to forklift everything out and then move all new equipment in. It's as it makes sense. >> As organizations have pivoted multiple times in the last 15, 16 months, as we've all seen, and there will continue to be that I mentioned, there's this sort of work from anywhere hybrid model, what are some of the benefits that a business could expect going from 100 Gig to 400 besides just quadrupling the speed? Talk to me about some of the business impact that can be made here. >> So business impact as is can be tremendous. Certainly, the capacity is the biggest one that jumps out at us here, as we can just combine, add on more services. Another area where we see this impact, and which, again, boils down to capacity, is IoT and edge. We have these new edge devices coming left and right. I mean, every time you turn around in the consumer world, there's some new thing that we never thought was possible, or we thought was 20 years down the road, and well, there it is. All of those cute little gadgets are just creating these streams of data, okay? So you just have so much more data that has to be processed. And so some of that gets processed at the edge, and that's kind of a cool new thing, but you still have more data that has to come back to the home base, either for storage, or for analytics, or for number-crunching, and so you have to be able to manage that. Bigger, fatter pipes going long distances, going short distances, going just in the same rank. >> Have you noticed, Jim, in the last year or so any industries in particular that are really prime candidates for this upgrade? As you mentioned, IoT, the explosion at the edge, sensors, sensors everywhere. Any industries that you saw that really are benefiting from doing this migration? >> Well, certainly the hyperscalers. The big companies that we all use social networking on. They're just moving around just piles of data, and everyone's working from home, and so they just have a little extra time to do the clicking and searching and stuff. And so that, and as well as entertainment. From home, people are just... They're just using up more bandwidth, and so the tier one, tier two providers certainly... We've seen just tremendous interest and growth as they have stepped out and adopted. >> Jim, can we do a double-click now on some deployment options and capabilities, maybe helping us understand it by industry segment? >> Yes, absolutely. And so some of the segments that we've been working closely with over the last 18 months here is like cloud service providers. Also large enterprise companies who have the large data centers. And then thirdly, federal is moving along very quickly. Federal's got all the security stuff that's been in the news of late. They have more calculation and just data transfer needs than ever, and so those are a couple of good ones. >> Got it, yeah. Ransomware is now, unfortunately, one of those common household words, as is pandemic and Pfizer, right? Talk to me about where automation comes into play as organizations look to migrate to become faster, to be able to manage more data coming in faster from more sources, where does automation factor in the mix? >> One of my favorite questions, actually, because in the networking industry, it has changed so much in the last five years. It used to be that when you were talking about large data centers, and just massive amounts of data, that the entire discussion revolved around these large modular chassis. And the reality is that nowadays, yeah, large modular chassis still exist, and they have a place, but they're not mandatory in all circumstances. And one of the big changes is that you can get building blocks that push out tremendous amounts of data within a single box. And you can use like a claw structure that allows you to do more data safer because you have higher availability than these really expensive modular chassis. And so when you come with kind of more switches, the reality is that now you have a bigger automation requirement. And so the tools to be able to automatically set that up, automatically maintain, and automatically monitor it, those are critical. And especially when we're talking about high capacity environments where you have millions of people watching the video being on the screen right now. It better be there no matter what blip happens on the backend. >> Yep. There's always that demanding consumer, (chuckles) no matter what you do. What about automating day one and day two operations? How does it play into managing this infrastructure, this modern data network infrastructure, with on-prem and in the cloud? >> Yes. So I work for Dell, and I forgot to mention upfront, I apologize, I'm a Dell employee, but I'm actually speaking from my opinion. I'm not representing Dell in terms of their viewpoint of all of these things we're talking about today. But one of the big things is that, as we have gone from those modular chassis to more individual units to get this cleaner deployment, the day one has to do with how do you design that. How do you, when you have more fiber cables connecting things up, how do you make sure that you don't oops, plug one into the wrong place? And so tools such as in Dell, we have tools like the Fabric Design Center that automatically generate all of those wiring diagrams for you, all of the testing. When you plug it in for the first time, it actually verifies that everything's clean. And then day two is monitoring what's happening. Are you getting issues, subtle issues that are maybe not noticed but are building up? And so things like the Smart Fabric Director can allow us to monitor those types of things and make recommendations for, "Hey, there's something happening we need to be aware of and watch it," or "Here's some corrective action." And so those kinds of tools really are the lifeblood to make sure that the team doesn't just get overwhelmed. And the reality is we all know as time goes on, we need, or we're given the opportunity to have fewer people working on maintaining stuff. And so you need more equipment that's more complex, but you have less number of eyes on what's going on, and so the tools just have to be locked in. >> So tools you mentioned. What about operating systems? Anything that you would recommend that customers looking into? >> That's another great question. So operating systems have changed. If you look back on the server world, you go back 20, 25 years ago, every server company, they made their own CPU, they made their own operating system, and then it evolved so that there was, now you buy a CPU from either maybe Intel, maybe AMD. But it's not like Dell goes out and makes its own CPU. We buy from other established leaders. When it comes to operating systems on the server side, the same thing happened. Well, the networking world has been catching up for quite a while, and so four years ago, we started talking about open networking, and the fact that there are options. You're not locked into just what is our primary operating system. And so there are opensource operating systems that you can run. There are things like SONiC, which has just really been taking the networking world by storm. And so we certainly support Dell Enterprise SONiC on our platforms. And that is another fantastic option. >> Excellent. Last question, Jim, for you. If you had a crystal ball, given the dynamics of the world today and how quickly things are changing, and how organizations need to be competitive, what are some of the things that you think we're going to see in the networking world in the next 12 to 18 months? >> Well, it doesn't take a whole lot of a crystal ball. We just follow the standards, bodies. We see that 400 Gig has really come on strong. And honestly, we played catch up in that industry, having all of the optics that we needed. We needed all the breakout optics to go from 400 to four by 100. Those took a good, well, six to eight months before those really came on board. And so now we're finally at the place where we're in a good place, but the next thing clearly is everything doubles. And so now we'll jump to 800 Gig over the same infrastructure, and so that's, again, everything doubles. And then there's a lot of talk about, "Well, what happens after that?" Well, then you go everything from 800 to 1.6T over that same infrastructure, and so it's just kind of mind-boggling capacity, but it's coming at us like a freight train. >> It is like a freight train. We'll say a good freight train. Jim, thank you so much for joining me on theCUBE today, talking to me about modern data center networks, what's going on there, the opportunities for businesses in any industry to take advantage of the latest and greatest. We appreciate your time. >> You bet. Thank you for inviting me. >> For Jim Wynia, I'm Lisa Martin. You're watching this CUBE Conversation. (smooth music)