 Live from Las Vegas, Nevada, it's theCUBE. Covering Knowledge 15, brought to you by ServiceNow. Okay, welcome back everyone. We are here day three, rapping down, getting ready to wrap down day three. Close out the ServiceNow live in Las Vegas with ServiceNow's Knowledge 15, hashtag No15. Join the conversation, browse the conversation, go to crowdchat.net slash no15 where we have a threaded conversation, CrowdChat there. This is theCUBE, our flagship program. We're going to go out to the event, let's distract the people in the noise, I'm John Furrier. My co-host Dave Vellante, our next guest is Alan Linewall, CTO Cloud Platform and Infrastructure at ServiceNow. Welcome back to theCUBE. Thank you very much, guys. So how do you feel? I mean, winding down. I mean, everyone at ServiceNow we talked to has had customer meetings, a lot of labs, a lot of action, a lot of tech geekness going on here, a lot of goodness in the cloud. That's your wheelhouse. Let's get into it. Yeah, it's absolutely my wheelhouse. I mean, this has been an amazing event. I've been here since Saturday and it seems like the minute I landed in Las Vegas, there's been customers asking me about the cloud, they've been talking to me about scaling up, doing new applications for service management, it's just been a load of fun. So I got to ask you, one of the things that we always cover, obviously all the Amazon shows, all the cloud shows, OpenStack, you name, we all cover all, and cloud agile development, DevOps, it's been around for a while, it's the rage and all the web scale companies and you guys employ a lot of folks who worked at some of those early companies that brought out this huge scale up open source communities is now hitting mainstream. You see all the startups start in the cloud called born in the cloud, but you guys have an enterprise footprint and know the enterprise grade game. That's right. So now you got the cloud of agility, you got the cloud as an innovation engine. Explain to the folks out there, what is that dynamic? Because it's not obvious to folks that are out there in the enterprise how powerful a cloud can be. Well, they know it's powerful, but when you put it into practice, to have the best of enterprise grade, land and expand, rapid time to value, standing up solutions. That's right. It's kind of your wheelhouse. So talk about that dynamic. What is it about your platform that makes it cloud native for apps? Exactly. I mean, what's really interesting about our environment is we really focused on building solutions for the end user. In other words, we want to give them an environment that allows them to develop apps quickly. We want to be able to scale that app very quickly to match almost any user demand and we want to be able to have that application and that service management portfolio sort of reach anything in the enterprise. You're right. It's a lot different than starting with AWS where you basically VI a blank file and you start writing your code from scratch. In our mode, we have forms, we have applications, we have integrations, we have all sorts of stuff pre-built that allows you to build service management apps really, really quickly. And because it is enterprise grade with redundant fibers and redundant power and lots and lots of big iron in terms of server capability, we have the ability to really grow as fast as the enterprise can grow. Talk about the challenges of operationalizing the cloud. You guys have a lot of customers that have been successful. They started an IT and the notion of service management isn't an IT game anymore. So you're starting to see encroachment into new territory, HR, financial ops, all these other cool stuff's happening where the value is, but you don't have to be, you guys don't have to have the full blown platform because it's kind of extensible. Explain what that's all about. In terms of that kind of platform, what is it about the technology that makes it work? Because we're here in buzzwords like real time, well, they're good buzzwords, real time, presence, private instances for developers. What, explain that. Yeah, so what we've done is we've built what we call multi-incense architecture. And what that really means is every customer gets a full stack of capabilities. Whether it's logical code running in a Java virtual machine or it's their own database presence that's in our infrastructure. And that does give them the real time aspect. It gives them the ability to add code on the fly. It gives them the ability to work in it, like you said, a DevOps environment where you can change things and immediately promote them into production. But another thing about the DevOps environment which is interesting is people like to be able to develop in one particular area and then promote up to production. And because of the fact you have multiple instances, you can actually do that. You can have developer instances, you can have instances for what we call UAT or user acceptance testing. And then once you're done throwing testing all that and having the folks in the user community with the enterprise play with it, you can move it into production very, very quickly. It's a product we have called Cloning. And it probably wouldn't surprise you that we do it with service now. So one of the things that we noticed for some of the developers, well, you guys have a lot of happy customers, service now, all the customers are happy, but the developers on the Amazon side who are happy love the fact that when they were pushed off the production, there's a lot of auto scaling, auto provisioning, auto updates on code revisions. What is the equivalent analog to service now? So if I move something into production, what does that mean to my code base? Does updates ripple through? How does that just explain that? Cause I never really got the explanation. Yeah, it's a bit of a different environment. I mean, our code base by definition has the ability to accept lots of changes. You can push new changes into our code basis, something called update set. And once it's actually in the environment, the environment itself does scale as needed. So we have the ability to basically move the customer instance we call it and actually scale it up dynamically. We monitor everything that's going on in the environment. We monitor the transaction rates. We see what the size of the incident tables are. We see how many users are coming in. And yeah, we will pick their instance up and we'll move them and we'll scale them and we'll make them bigger. They don't even have to ask us. It just sort of happens on demand. And the reason that's so powerful is we don't want the enterprise to think about, I need to now think about a number of CPUs I'm going to add, or amount of memory I'm going to add, or we just want to think about deploy the application and we'll go ahead and scale it automatically for them. So Dan McGee was on earlier and he's the top dog in terms of all the infrastructure to build out. And so I asked him, I said, you know, what's your plan? I mean, people don't look at service now as a cloud company. They look at Google, Azure, AWS, but you guys have a full blown cloud. So explain that out. And two, talk about the geographic data centers. Are you guys global? You guys okay in all the countries? Sure. Tease that out. So how about the cloud? What's going on for you guys in terms of having that cloud capability? Sure. So we actually have 16 global data center presences. And what we do is we deploy our cloud in pairs of data centers. So the US we have a pair, Switzerland we have a pair, Canada we have a pair. And between those pairs, customer instances have an active side and sort of a passive standby side. And the idea is in a given geography, you want half your actives on side A and half your actives on side B and they're copying data to each other dynamically and keeping them all up to date and up to sync. So our cloud presence is really geographically based and it's global. It's in 16 locations and eight different geographies. And those are synchronous locations? That's right. Okay so and you're saying you've essentially got data in two places. That's right. But I was confused, you said one's active, one's passive, you said no that's a typical situation. Sorry, what happens is for a given customer instance, their active instance will be in say side A. But at the exact same time we're replicating their data in real time over to the passive side and side B. And that gives us the ability that if something goes bump in the night and side A, we can quickly switch them over to side B. And that's what we do. And so how do you deal with sort of, you know, RPO around or how do you deal with local disaster, for example. Hurricanes, earthquakes and things like that. Yeah, we have the ability to move customer instances. Matter of fact, there's probably 3, 400 customer instances being moved between the active and the passive side right now. The reason it is is because we have it all automated and we're always trying to balance our cloud. So let's say we're talking about Switzerland. In Switzerland we have a portion of our cloud Geneva and a portion of our cloud in Zurich. And half our customers are active in Geneva and half of them are active in Zurich. And the backup side is on either side. And again, at any point in time, if we wanted to, we could pick everyone up from Zurich and move them to Geneva. But that's in a synchronous distance, right? So how do you protect against, you know, some kind of disaster within that region? So if all Switzerland went away? Yeah, well, I mean, it depends on where your stuff is. And Switzerland's not a flood zone. Yeah, so in the United States, we have Virginia and we have San Jose. And we try to make sure our data center pairs are in outside of sort of hazardous zones. So they're not paired in the same location. Okay, but so then that instance is not synchronous. It's not synchronous distance, I mean, 200K. Oh, I see what you're saying. Yeah, so you've got an RPO data loss factor that you've got to account. It's about 100 milliseconds max between any two locations. Okay, okay, so that answers the question. I mean, you're essentially dealing with that with distance. Yeah, so I thought you meant by synchronous copy. Okay, I'm sorry, yeah. So in Europe, it's very common to have physically, I mean, that's sort of the norm. United States is obviously different. Yeah, and in Europe, we run in Amsterdam and London, and then we run in Zurich and Geneva. Okay, so will customers in Switzerland actually have data in London as well or no? No. No, okay, so you keep it in? That's right. I see. And then I wanted to ask you, because we're going back to the architecture. You talked about sort of Amazon, you get basically a blank piece of paper and away you go. You guys are providing primitives as part of what you're delivering. We're also providing full applications that people can leverage and modify and tweak on their own. So you've got essentially IT primitives and you've got business primitives, is that so, what are some of those sort of concepts that we should be familiar with? It depends on which discipline you're talking about. If you're talking about an IT discipline, we have your classic incident problem change, which customers that can tweak for their own environment. Data structures. Data structures, forms, lists, workflows. Something becomes an incident, an incident could become a problem or a bug, a bug could mean you have to implement a change. But you can do things like facilities. Facilities, you might want to have an incident which says the restroom's dirty. So you want to file an incident to say the restroom's dirty. The way you do that is the changes, someone gender comes in, fixes it and changes the restroom. It's kind of the same process, whether you're talking about IT or facilities or HR. Think about from a HR perspective. You call HR and say, I want to know what my benefits are. That could be an incident. Somebody called up and said, tell me what's going on on life or death benefits. And then the way you solve that is you have someone give them form back, or have them give you information back. No, out of the box, I've got business primitives too. It's your workflow oriented, business process oriented. Exactly right. Hey, well you brought it. That's got triggered me for some good questions there to ask because I love this stuff. Part of my job. Don't throw any water on me. But you mentioned lists, data structures. I'll see you get databases in there. Real time's huge, presentation, presence, real time, synchronous, client, type of edge. You got Angular, you got node stuff in there. All this stuff's happening, right? It's the modern way. I think it's managing cameras and all the stuff on the devices. Really brilliant, great job. I really love the software. But that brings up the question of unstructured data because now you're dealing with a new dynamic so I want you to share kind of the complexity of really how hard it is to do that real time stuff. When you're dealing with databases that have to sometimes keep their stores and or you've got lists. Things like Redis have been successful out there. You've seen a lot of things like streaming in the cloud. What's under the hood there? Can you just talk about what you guys have to address that because the developers need to have queuing, they need to have lists and they need to also have the standard databases in there all working at low latency real time. What's the secret sauce? What's the magic under the cover? It's interesting. I mean you're absolutely right that developers do need that environment but in our environment we don't necessarily have those specific data stream primitives for example. They're built into the platform but specifically what we have at the top end is we have a horizontally scalable array of Java virtual machines. We're running Tomcat. So all the application logic runs in Tomcat. So the presence is done by having that code in the Tomcat Java virtual machine. There's some other present servers that sit off the side that understand who's online and when they're not online. And then behind all those Tomcat servers there's a bunch of MySQL databases. Again a MySQL instance for every single customer. In terms of document store and things like that we actually handle most of that by logic in the Java virtual machine and we store in the MySQL database what you might consider a blob, a data blob. But that's a natural decision, right? Architecturally that makes a lot of sense because the Java you decouple is application driven. And then that's decoupled, highly cohesive in itself from the real time stuff which is dynamic and always moving. So you build that into the platform. We do, we build that right into the platform. We also have queuing mechanisms and other technologies in the platform that allow us to keep those things up to date. We don't necessarily expose them out to the user. We expose that to the user is an API that allows them to benefit from that service. Upload this particular photo. Find out the presence of this particular user. They don't need to know about looking at a particular queue and registering for events. It's just there in the API for them to use. And this, I go down this line because the question, because this is one of the things we heard all week is that it makes DevOps not DevOps. It makes it, you abstract away a lot of those complexities. So now you're essentially just filling out a template. Not, I don't want to oversimplify but you know what I'm saying. You don't need to go in and run the queuing and you can use the queuing. You can just use the existing predefined. Technology, you've abstracted that. Is that correct? That's right. I mean, what is DevOps? It's operations and development working together. And it's development trying to solve the business problem and it's operations trying to keep infrastructure and owning that entire stack and simultaneously those guys working together. Previously, as you know, developers run something and throw it over the wall of operations and say, you know, you're piled, they deal with now, go figure it out. And we've sort of changed all that now where people can develop in real time and the operational mechanics of everything behind that is kept up to date automatically. Scaling, for instance, is an inherent, twiddling around with elastic beanstalk. No, you're not twiddling on elastic beanstalk. This is the holy grail, in my opinion, pun intended, because what you have is the DevOps, what I call Gen 1 DevOps, so the guys who built it themselves, they were writing the code. That's right. It tends to our early comment that Fred was on, saying that you guys hired a lot of guys who did that, making great architectural decisions. But actually, focus isn't from pure developer engineering to business engineering. So this is where the focus and the service comes in. So what's interesting is you've abstracted away, provided a platform for folks to actually work on the business process engineering, which includes writing some code. There's definitely, I mean, there's definitely, if you want to customize your application to read a particular medical record automatically and perform OCR on it, to have it fill out a form, you know, we don't have that. But you could write that code onto the platform and the platform would use the basic primitives in the API and the REST API to push that right into the incident table or to push that attachment straight into our system. And then on the back end, we would record that in my SQL and we replicate it to the parasite. We make sure all the network traffic worked. You don't have to worry about that anymore. Yeah, this is why I think the big shift is good because you're standing stuff up. That's right. The word standing up is the cloud term. Where does that, but that's the new normal, right? Standing stuff up fast doesn't have to be just pure tech. It could be the business loose and our coding. So that is, to me, you guys really, I think have a secret formula. They're lightning in a bottle. Which brings me to my next question, developers. Yeah. 1,200 developers here today for creating, isn't it? Pretty significant traction. I mean, pretty big traction. So where do you guys go from here? So I got to ask you. You're providing essentially the engine for whatever cars people build, and metaphorically speaking, for these applications. What is the continued approach for the developers? How are you guys building out the infrastructure to create more goodness for these developers? Let's give us some tease of what's coming around the corner, what's out there now. I'll see private instances we mentioned earlier. Yeah. Some sandboxing capability. Yeah, we have this capability called developer instances which we're giving away to our developer community. And the idea is that we want people to get into the platform. We want them to be creative. We want them to be able to take the building blocks of everything we built and to sort of run with it. Now, our internal folks, our applications, whether it's our service management applications, our HR applications, our ITOM applications, they already take advantage of that. So I kind of have two users, if you will, with two customer bases. One of them is our internal application developers. We've been getting free instances for a long time. We've been getting a lot of them. Depends on how you define free. But now you're exposing free instances to the real world, the rest of the world, right? That's right. What else do they get with the instances? Just instances, they get any more code in there from the applications? What are they? Well, they can go to share and they can download things that the community has given them. They can obviously buy other things that people put on the store, which we just announced. And they get a whole set of functionality that allows them to create almost anything. Go walk the X-Fo now, show floor. It's just amazing what we've seen people do. We saw one team develop an event correlation tool that looked into CMDB. And then I turn the aisle and I see someone that's working on medical and legal records by writing code. It's just such a diverse environment. I ride in the elevator and I see folks from Corning and I see folks from Arthur Anderson, I see folks from Ernest and Young and Morgan Stanley. And I'm like, these are just all different verticals but they understand that what we've built in the platform allows developers to really sort of go wild. And what we've done with CreatorCom has really given them that vehicle where they can get together and lock mines. Yeah, Dave and I were talking earlier about why investors should be excited about servers. Now we're talking about the whole value proposition. One of the things that I didn't mention that we highlight here is that it's for all businesses. It's not like an industry. So Morgan Stanley, that's a melting pot, Morgan Stanley. So that's awesome. So I got to ask you, pretend for a minute that we want to port our application, the cube to written in Node, Angular and Java. We have three separate code bases. What would we do? How would we, what would we do? So I come back to the outside, hey, just came back from service now, amazing platform, they have a store. We can make money with these guys and solve our customer problems without doing all this other stuff. Take me through, what's that next meeting? Do I got to start writing code? I got to get the free, get the instance. What's involved under the hood? Do I do a language, do I code? Yeah, you get a free developer instance. And then you start learning our glide APIs and you start understanding how to interact with the platform. I don't know exactly what your application does but I assume it's some sort of registration service as well as some other service people can watch videos, chat, well chat's already there. So as soon as you stand up and app on the platform you've got a chat room. So there is no code to be written for chat. If you want to do a live feed where you decide that you want to post messages and have everyone on the portal see you live sort of feeding and what's going on, that's there, you don't have to do anything. It's already there. So this is the sort of benefit of the platform. One of the things you'd have to write is you'd have to customize any sort of interaction with those users to meet your needs. Does that make sense? Glide APIs mainly, that's the main thing. Yep, exactly. And then I write my app for my world, solves the problems and I'm interfacing with those APIs. That's right. And that's it, pretty much. Glide isn't Glidesoft? Glide isn't Glidesoft, that's right. He's not giving up on Glide. The founder, he got his little walk away. He got his sacred cow in here. The internal code still is called Glide and people ask us what it's about and we just say, there's a history there. You got to keep the sacred cow in there. So API good, documentation will be there. Know the videos are on the site and then ultimately certification for the store. That's right, you could end up becoming certified for the store. It's funny, I ran the hackathon on Tuesday night and to see the things that people wrote on the platform. We had a team there that had been on the platform for 48 hours and they were writing an application to help Toyota Motors that work for Toyota to fill up an inventory management system to make cars. So when you order something specific on a particular type of car, they wanted to be able to create an order form that started a workflow that told the workers on the manufacturing floor, go put in tan carpet in gray seats or whatever you want to do. These folks have been on the platform for 48 hours and they're going to cure that app in less than eight hours. And that's because of the platform tool and the primitives you provide, all the distraction you've taken away from them, they're focusing on the front end. They're front end, the front end and being sort of the infrastructure guy, the thing I'd bring up is once we stood that up and put it in production, it would be in two locations, it would be accessible via anyone, have our full security parameters and perimeter around it. It'd be replicated, it'd be backed up, it'd be available. It would be, you know, if anything on that particular hardware went south, it would move without them touching it. All that's there day one. So what's the coolest thing you've seen here at the show and I want two answers here. One, in the show itself, keynote, labs, whatever, and two, the hackathon besides the Toyota examples, give some other examples. Okay, the coolest thing I've seen here at the show so far, I gotta tell you, Fred's demo was pretty cool. We all were sort of holding our breath knowing what was going on behind the scenes there, but that team always just pulls a miracle off. So we were really impressed by that. That's a little self-serving, but that is sort of the thing that I kind of went, I want the eye watch now. In terms of the hackathon, I think the coolest thing I saw is we had a team from a company called GoDaddy. GoDaddy uses us as our IT department. And there's a team of five people and they're trying to build an application that said we have a platform application that allows you to have on-call schedulers. So what that means is, in my world, I have a bunch of database administrators or a bunch of system administrators and you can say who's on-call for which particular nights and times a day. And they use that, but the thing that wasn't available is if I'm driving, take me off the schedule because I don't want to be talking and driving on a problem, or if I'm out at dinner with my wife, hit a button and say I'm out at dinner with my wife, take me off the schedule for two hours. So they had a team of four and they started working on it at four o'clock in the afternoon and by 7.30 at night, three of the four left. They said this is too hard, we can't figure this out, we don't quite get it. One woman named Anna Stade. She got all the way to midnight and actually won one of the five golden tickets to go on to the next day. She took it home solo. She took it home solo. And again, she had never developed on the platform before and by the time she was up on the Expo Now floor, you could click in and out of your on-call rotation and it's all developed. Well, just Alan, I think he's going to have big success with developer programs. So make sure we bring theCUBE to those events because I think you guys are going to have a great business model, how you built the platform, not being too aggressive on the land, grab side. Got great enablement, you're empowering people to be successful and people make money and have fun doing it. So congratulations. Thanks for spending some time on theCUBE. We really appreciate it. I'll give you the final word as we break. What's the vibe of the show? The folks aren't here. Share with them out there. What's the vibe of the show? What's it like? They're not in the moment like we are. I know you're working hard, but there's some good vibe here. It's fun. It is. It's a lot of fun. It's a lot of fun. There's a lot of folks coming out from CreatorCon right now. I think the thing that's most amazing to me is just the scale of this show. I mean, there are people everywhere. We were talking just a few minutes ago. There's people out on the lawn. There's people upstairs in the Hang Now area. The community is buzzing. Everywhere I go, people are like, have you heard about this? And look at this particular app, and we're going to go home and use this for X, Y, Z. We're going to come back a year from now. There's going to be 10,000 things that have been developed that we've never thought of. And that's just awesome. And that's the whole purpose of the platform. Really, congratulations. I really, a big fan of what you guys done. Software is phenomenal. The demo Fred gave us. And it's really hard to do. And it's great software. Again, product leadership is one of the things you guys hang your head on. Yeah, we're having a lot of fun. And I think everyone on the team just really enjoys providing solutions. And the feedback we get from our customers is just overwhelming. Yeah, you guys are very humble. And you take the feedback. That's great. Some companies don't. I'll name a few, but I won't. We love feedback, because you win us water in our face. But that's theCUBE. Allen Line 1, thanks for coming to CTO, Cloud and Infrastructure here inside theCUBE. Sharing the signal from the noise. We are theCUBE. We'll be right back after this short break.