>> Announcer: Live from Las Vegas, it's The Cube. Covering AWS re:Invent 2017. Presented by AWS, Intel, and our ecosystem of partners. >> Welcome back to The Cube's continuing coverage of AWS re:Invent 2017. We're live from Las Vegas, day three. There's still a ton of people here. We have a great guest next and we're excited to talk about Mark Kelly, the director of Cloud and Infrastructure Services Architecture, talking with myself, Lisa Martin, and my co-host Stu Miniman. Mark, welcome to The Cube. >> Thank you, glad to be here. >> So Scripps Networks, tell us a little bit about that. I know a few things, HGTV, Food Network, I watch those a lot but, tell us a little bit more about Scripps Network. >> So Scripps, when I say that name, most people do not know it. But we are the leading provider of home, food, and travel content for broadcast, the web, and for emerging technologies at this point in time. So, we've got a lot of brands in our umbrella. You said HGTV, we've got Food Network in there, Cooking Channel, DIY Network, Great American Country, and Travel Channel. >> Yeah so Mark, I know my family binges a lot of these shows so-- >> Really? >> As do I. >> Take us little bit, you know, what's happening in that industry today? You know binging versus watching online, cloud, digital, the joke of course is that at least things are pretty stable and not changing in your world, right? >> Oh I know, absolutely. It is changing on a very rapid pace so, as more and more people, you know, the big buzzword is that cord cutting thing. We have got big concern in out industry around that. But as more and more people talk about it, we're adapting our technologies to be there just to provide for them. So we're on those emerging technologies. We're on all of the said top boxes. Apple TV, Roku, we're actually developing networks specifically for those technologies. So we're trying to adapt. The broadcast world is still our bread and butter, so until we figure out the revenue models for these new technologies, we're gonna be in the broadcast world for some time. >> So Mark, talk us through a little bit of what's your role there. You've been there six years I believe you said. How has Cloud changed the way Scripps works. >> So, Cloud has definitely changed the way we work. When I first came on board, Scripps was looking for a way to move the business faster. Scripps, because it is changing so rapidly, they're in the business of-- They have to compete with startups. And when they're competing with startups, startups don't have the same process as it controls around everything that the bigger enterprises do. So it was an extreme challenge to them. And you know, when they don't understand-- They don't have budget concerns like big enterprises either. When I go in to do a project to launch a new network, they don't understand it. I've got to buy hundreds of thousand dollars worth of hardware or whatever associated with it. So when we're looking at the Cloud, we needed a way to accelerate that. We wanted the make the business process much quicker. We weren't going to have those lead times for purchasing hardware, purchasing software, getting licensed, you know, racking it, stacking it, doing all that. We were looking to ways to move that out of the way so everybody could focus on providing more value for the business. And that was the primary reason we actually started looking at the cloud. >> So from an end user goal perspective, obviously you mentioned speed and having to compete with other networks and other native original content companies. How does the end users demand for 24 by 7 content, how does that drive really, the pace of innovation that Scripps has to meet? >> Yeah that is a very good question. I think we're still trying to figure that out because the whole consumption method for all of our end users is changing. They're going from their traditional TV to where they're watching it on their phone, to switching to their iPads, to everything else in between. So we're developing technology methods so that we're providing it for them as they need it and want it. >> Yeah, Mark can you give us a snapshot of what is your infrastructure look like today? What are the kinds of major things that have changed and what's kind of on the table? >> Today we are running about 3000 instances in AWS. So AWS is our primary cloud provider. We run a lot of auto scaling to take into account our seasonal loads. That's probably one thing that we-- One of the challenges we had going into the cloud is that we were up against certain times of the year when the load was extremely high but it was only extremely high for a couple of weeks. I mean literally we'll be running at two to three hundred percent of what normal was, and it would do it for two to three weeks. So we were having to purchase hardware in advance for that and it would just sit there idle for the rest of the year. So we were looking to ways to avoid that excess purchase and still scale to support our consumers and needs. >> So when you started this journey, and we were talking to Veeam earlier about-- Cloud is a destination, you get there, great, you're done. But when you start this journey you mentioned AWS and we're here at re:Invent, where did you start? What were some of the stakeholders that you had (Mark laughs) on the business side that you had to say, >> Absolutely. >> Alright guys, let's come to the table, we have a great opportunity here, this industry is transforming, as you know, the joke Stu made earlier. Where did you start with those business discussions? >> So I think it was more around trying to figure out that solution. It's like when I said earlier we're trying to get to market quicker. So it was easier to go to the stakeholder and say, "Hey I think I can go this route, and I think I can get you to market twice as fast." >> Lisa: Their ears perk up? >> Oh absolutely. Ears will perk up, eyes will light up. They're like, "Really? How much more is this gonna cost me?" And I'm like, I don't think we're gonna-- They went to represent the on-demand cost models. I think we don't have to do all these big upfront costs and that got a little bit more excitement out of the stakeholders. And we went forward from there. So we would um-- It took a while. Don't get me wrong. Everybody gets set in their ways. They're going to do things normally. Change sometimes is difficult to push through. >> Lisa: Absolute-- Cultural change. Very, very challenging. >> Absolutely. So it is a complete cultural change. They go from your traditional on-premise to the cloud model where you're not managing the hardware as much. You're focusing on the engineering around it. >> So Mark, you bring up managing. You know, monitoring something that's gonna change very differently. Talk a little bit about your people, you know, the skill sets they had to change, what's the difference is kind of before, during, and after that migration. >> So beforehand, when we were on our on premise environment, we were focused on tool sets, for monitoring, and managing infrastructure that were very vast. I mean we had a large number of them. One of our goals moving to the cloud was to consolidate that tool so that we wanted to get down to a minimal set that would actually accomplish our goals to get everything that we needed. So we didn't have to go through training people to learn 50 different tool sets for monitoring whatever it was. Network equipment, storage equipment, or the computer equipment. We were actually focusing on monitoring our applications and still getting some of that underlying infrastructure reporting on our monitoring, but we didn't have to have the same level where we were monitoring the hypervisors, we were monitoring the network switches, that kind of went away. So our focus became more on the operating system up and engineers developed. They no longer had to focus on that hardware. >> So what are you using for that, you know, how many people does it take to do that? Do you know if it can kind of compare to what you had before, yeah. >> Our standard monitoring switch is actually composed of New Relic. New Relic has been a great partner for us. They had the same mindset as us. We're looking to compete with the startups, at the time New Relic was the startup. Because we've been using them for between five and six years now. We brought them in because they had that same hunger and mentality that we were looking for. Their culture mashed really well. And we got in and deployed all of their suites to every environment. We actually leverage it in from development up through production. So we try not to separate our monitoring. We try to keep it all uniform so out troubleshooting gets a lot simpler. We actually have the same people monitoring our DeckStats that are monitoring out production stats. So they can troubleshoot and help get through there. It became a lot easier for them to do that. >> Could you talk about kind of, how many people you had, managing infrastructure before versus the monitoring, and some of the training they had to go through or? >> Before, it was all specialized people. I don't even have that. I couldn't even give you the headcount cause most of that was before my time. >> Stu: Yeah. >> But every individual was actually specialized on each application for monitoring. So it was actually teams that would focus on each part of the business. So as we migrated into cloud, we became more standardized. Made it simpler. We actually have our-- My cloud team is really in between about 15 to 20 people. And we're managing 3000 instances in AWS, and about six petabytes of storage. We've got quite a bit of a content up there already. >> And you're on the customer advisory board for New Relic. >> Mark: I am. >> You mentioned cultural alignment between Scripps and that vendor, and that's really key, but talk to us about this collaboration and it sounds-- Maybe by direction of that, you're able to maybe influence some of the things and help them make their technologies better. >> Absolutely, we've got a really good relationship with them so anytime we have a challenge, one of our current challenges is serverless. As we move, we have a lot of development teams that want to move into serverless. We've been working with the New Relic teams and giving feedback to them on what our challenges are with that, and how we're monitoring it, cause we've got certain things where I wanna be able to monitor those functions in serverless, and I'll need to give a cost back to my stakeholders to say this is what you'd cost. It's challenging to do that now but we're working with the New Relic team to help them deliver some of that knowledge to us. >> Mark, you hit a hot button for me. So bring us inside, you know, why serverless, and what are you hoping to gain from that. I've seen New Relic actually has been tracking for the last couple of years, adoption of containers in serverless so-- >> Containers in serverless is kind of the new hotness you know. We've been moving into serverless primarily because again, it's the next generation of speed for us. Makes it even simpler for the developers to get started, we can give them a standard framework, they can start developing their code and just push deploy and it's running, and they don't have to worry about any infrastructure or managing anything. Again, the challenge has been the monitoring part of it, but working through that and actually getting pretty good results out of it so far. >> So you've got about 70% of your consumer facing side is on AWS which has got some latency sensitive workloads that are still on-prem? >> A hundred percent of my consumer facing properties are on AWS. >> Oh fantastic. >> We do have some workloads that-- Those are really not designed for cloud. It's our end use, end financial systems, our critical business systems that need to be close to those departments. Those actually still live on-premise for us. When we started this journey, the on-premise was-- it was a slow, horrible process but, as we evolved, the cloud, they've evolved that on-premise stuff to keep us with them as well. We're actually looking at, so you know, some of the other monitoring solutions out there. New Relic has been an option for us to actually look at on-premise will, monitoring as well. >> So all the advancements that you guys have achieved in your six year or so transition to cloud that you've talked about. What's next for Scripps? What are some of the maybe new business opportunities that this optimization, cost reduction is enabling? >> So next for us is actually machine learning and AI. We have large initiatives going on that right now. We are trying to analyze our video, analyze our content, lots of it's to help remove some of the manual processes that we have now. Because a lot of that stuff when you're delivering to our different partners, there's certain requirements around the video, and the only way to do it right now is with eyeballs watching the videos. So is this somebody sitting there watching it for hours and hours a day. We're leveraging the machine learning stuff to actually auto classify this video, pull out thumbnails for the authors so they can put it in there, the metadata awareness form, and we're doing lots of things with AI. So we're looking for that to be a really hot feature for us in the next couple of years. >> Excited with what you heard this week from AWS about AI in ML? >> Absolutely. The first day keynotes were completely blown away. They were all things we were looking for. >> Stu: Anything specific that you've been waiting for, or just not waiting for, but got excited by? >> Yeah there was lots of it. The Kinesis video streams were actually really good. The video API, I'm drawing a blank on the exact name of it, but that one actually had some really good features for us because we are looking to do exact things that that one does. We're looking to pull time stamps out for when stuff shows up in videos and provide that back to our end users where they can search and find things in the videos much more quickly. >> Excellent. Well Mark, thanks so much for stopping by The Cube and sharing what you guys have been doing at Scripps Network with us. >> Mark: No problem. >> It sounds like you've seen a massive transition and you're really have a great foundation to continue going forward and >> Thank you. >> Lisa: And we look forward to continue to watch great shows on the network. >> Awesome, thank you guys. >> And for my co-host Stu Miniman, I'm Lisa Martin. We would like to thank Mark for stopping by. You're watching The Cube's continuing coverage from Las Vegas of AWS re:Invent 2017. Stick around guys, we'll be right back.