 Live from Las Vegas, Nevada, it's theCUBE. Covering Knowledge 15, brought to you by ServiceNow. Okay, welcome back everyone. You are watching SiliconANGLE Wikibon's theCUBE, our flagship program, we go out to the events and extract the simple noise. I'm John Furrier. My co is Dave Vellante with wikibon.org and we're pleased to have Pat Casey, VP general manager of Create Now, platform development, early employee of ServiceNow, great perspective. We're going to get geeky here, but talk about some of the high level stuff. Welcome back to theCUBE. Thank you very much. So you've seen the evolution of ServiceNow from early days to public companies, scaling, very cloud. I mean, it's inside the tornado to use that metaphor. It's been so successful. What do you feel? What's your feeling right now? And how much more work do you see on the horizon? Well, I think probably the first thing I feel is shocked is the honest answer. When this company was founded, we didn't have office space. So we borrowed office space in the basement of our VC and it had no windows. So we're in this little tomb of a room and there were five people there and one table we got from IKEA. So to look out now, we've got 9,000 customers who paid money to attend an event about this. It's just, it's shocking. It's also humbling. And it's also to be honest, it's scary. People are here because they are dependent on technology that we wrote. And one of the things that's just been always been sunk into my head and I believe this for me is I do not want to let anybody here who has put their faith in ServiceNow down. So in terms of where the work is, we've only just gotten started. I get up every day and I am just, I fundamentally want to make sure that this is the best product it can be and our customers get the best experience. Great passion. I mean, that's a startup hatch, but you guys know I started your big company. But you got some good things going on to get some wind at your back to use the Frank Slutman sailing analogy. The market is exploding with innovation. So that's a challenge, but it's also could be a great opportunity. So what's your take on it? I mean, you got the agile, you got native, we're hearing terms like microservices being kicked around in this native cloud apps world. You got a platform. Share with your take on some of those buzzwords and some of the big mega trends. I think if you, when this company was founded, this is actually founded as a platform company, which I think most people don't realize. But when Fred sat down to design this, his cocktail napkin design, and there was actually no cocktail napkin, but imagine there was, it was we're going to run enterprise business apps in the cloud. That was the idea. And the first few sales calls, though, selling a platform were kind of miserable because we'd go to the customers and we'd say, hey, we're here to see, show you service now. And they say, well, what does it do? And we'd say, well, whatever you want it to do. And they kind of cock their head and say, well, it's your sales call, guys. You got to talk to us. So we built out a suite of applications on top of the platform. So we'd have something concrete to sell. And that's what the company sold for probably about eight years. It was our ITSM suite, incident management, problem management, change management. That's what most of our customer base uses. We're sort of pivoting back to focusing on the platform again, though, partly by building other apps. We've got HR, we've got facilities, we've got legal, we've got GRC, but it's also about trying to get people just onto the platform itself. And in terms of really big mega trends, that is one of the mega trends we're seeing. It's that people are not building everything from scratch anymore. It's just not an efficient way to build things in the market anymore. And people are also moving to more and more specialized pieces of tooling. You don't start with a C compiler anymore. You start with the higher level language. You start with Ruby on Rails. You start with J2EE if you're an enterprise developer. You pick a tool that's appropriate for the problem you want to solve. And service now is a great tool for solving a lot of enterprise business applications. Let's talk about developers because one of the things that I hear all the time is, oh, I built this on Node. I got this in Angular, I got this in Java. There's a lot of different stacks kind of being built, but cobbled together can, you know, I guess I'll put them in a container. Whatever they say these days, there's a lot of cool stuff happening on the developer front. Open sources, we're doing great. What are you guys looking at in terms of leverage? And oh, by the way, that enables non-programmers to do stuff that looks programmatic. So the innovation opportunity for creating is huge. So what's going on with you guys We actually view the developer world as kind of being in three different groups. You've got, it's a Gartner term, but I think it's a good term. You've got low-code developers. And that's someone they can make a form, they can make a list, they can potentially do a little bit of light scripting. It's your kind of traditional system administrator archetype. And that's who we founded the company to address. That was the business idea. We could enable low-code developers. We could enable administrators to build really meaningful business apps. And that's really been the secret to our success. We're really good at it. Because they're closer to the action, but don't have to go in and go out of band, if you will, to kind of develop requirements. I think most people do their best work when they're scratching their own itch. So if you're close to the problem, you're like, man, I can solve this for myself. And it's been very empowering to let administrators and low-code developers do that. But that's not the totality of people out there. There's also people who can't even do that. They're no-code developers. They're my mother. She can use Excel really well, but she can't write code. And my mom is a very bright woman. She's a healthcare consultant, but she's a no-code developer. But she can put a spreadsheet out there with column headings. She can make forms. Using our no-code tool, she can actually put a business service out on the web with approvals, workflows, notifications, dynamic forms. Matt Fiverr put out a HR app in one day when he started playing with Express. Absolutely. That's the trend, right? That is definitely one of the futures you see is this democratization of access to development tools. It used to be when I started in this industry, pretty much had to be an educated professional to build anything meaningful. That's no longer the case. You get kids today building great applications with real business value, real value. And that's the value of the modern era. The barrier to entry has just declined and declined and declined because the tools have gotten so much better and so much more specialized. The combination of the two is just incredibly empowering. So I wonder if we could talk about architecture. Maybe, I don't know, inside baseball or maybe plumbing. I don't know. But you said in your keynote, multi-tenant is the TV dinner of cloud vendor deployments. What did you mean? Let's talk about multi-tenant versus multi-instance. Sure. So traditionally in the SaaS space, there's really two different architectures people deploy. The most common is something called multi-tenant. And multi-tenant, if you imagine it's a big old apartment building where there's one big construct, there's one big database, some software on top of it. And each individual customer is a separate software construct. You're sharing hardware, you're sharing software, you're sharing memory, you're sharing an apartment in an apartment building. It's really sort of efficient for the vendor. It's certainly convenient for the vendor because they've got one thing to manage. You think about it though, there's downsides though where if the water main breaks, you have the entire apartment building or every customer in this case, they don't have water. So the failure modes tend to be really extreme with multi-tenant environments and you can't do things like let people paint their apartment any color they want to or expand their apartment or cook foods that are really smelly. You have to have apartment rules in place. And you see the same thing with multi-tenant architectures where in order to make it work, you have to restrict what people can do within your platform. You get licensing restrictions, you get technical limit restrictions, you get wrapped up in quotas. That's part and parcel for multi-tenancy. ServiceNow is not multi-tenant, we're multi-instance. So every time a customer joins us they get a unique instance of ServiceNow. It's just for them, it's your own house. And because of that, we don't have to go in and tell you what you can do with your house. There's no HOA, you can paint it green, you can paint it pink, you can do whatever you want to because it's yours. And that's the big freedom that we can do for the enterprise customer base, for big customers. And multi-tenancy does have its use case. I don't want to oversell it. If you're selling largely into kind of the SMB space for example, it's a really good architecture. But up at the enterprise level, it's really not. The multi-instance architecture we use is fundamentally, I think, superior. Okay, so what point did you make the decision to go to multi-instance? Obviously early on, you were there early on, and why did you make that decision? I think it's not as clear cut as it is in history. You always look back and say, well, we had this great design system. We set out knowing we wanted to address the enterprise space. And we eventually figured out that in order to do this we couldn't do it with multi-tenancy. But we had sort of talked ourselves into kind of our own little version of, I don't know if you ever watched South Park, but the underpants gnome dilemma. And if you remember that episode, Cartman and I think Butters, they decide they're gonna stake out the underpants gnomes who sneak into your house and they steal your underwear. And they follow them. They watch them steal some underwear and they follow them down to their underground lair and they accost them. And they say, why have you been stealing everybody's underwear? And so the gnomes take them to a small room and they show them PowerPoint. And the PowerPoint has three parts. In part one, the gnome steal underpants. And in part three, the gnome's profit. And then they skip back to part two and there's a big question mark. So we had the same problem. We knew we wanted to go with multi-instance and we knew it was gonna be great in the market. We had no idea how to do it. So we probably spent about three years of engineering effort figuring out how to make a multi-instance architecture work well at scale. Because doing it once, it's really easy. We have 18,000 instances in the platform right now. That's a lot. Things have to work with automation. They have to work cleanly. And they have to work all the time. So it wasn't a matter of convenience for you, just the opposite. Oh, absolutely. It was a terrible challenge. It was a challenge we had to overcome. And I think it was necessary for our target audience. And if you're listening to this and you're actually looking to start your own SaaS company, figure out who your SaaS audience is. If it's small business, if it's medium business, multi-tenancy may be absolutely the right answer. And the trade-off is cost, efficiency. I mean, it's more expensive, right? So. Not necessarily. I think there's this myth that it's more expensive. It's not convenient. You do do more engineering work, but in terms of what we actually spend on hardware and power and cooling the data center, compute is compute is compute. If I have to buy a lot of servers and plug them into one database, or I have a lot of servers plugged into a lot of databases, it generally equates to roughly the same hardware cost. So it doesn't generally drive CapEx, but what it does drive is you've got to put that engineering effort in. It's work. And you're not a data intensive app. I mean, you have a lot of data and servers now, but. If I remember my numbers right, we're about five petabytes of storage. So that's not. We are not saying. Netflix, you know, we are not box. You know, we're not storage centric. It's transaction. So you guys are optimized for transactions. Absolutely. But the implication that you've made is that many of the clouds that are out there are fine for SMV maybe. If you're an SMV that is okay with that, but many are not suitable for the enterprise. Absolutely. And I think that's the big change we're seeing in the cloud space. I'll use a different analogy, but a hundred years ago, just under half of all the cars on the road were one model, was the Ford Model T. It was like 48%. And the best selling car was actually a truck in 2014. It was a Ford F-150. It was 2.3% of the market. The day when one car could dominate the market like that has long since passed. But in the early days of the cloud, there were only a few vendors. So they were trying to address as much of the market as they possibly could. So they built very general case solutions. Well, time has changed. People are getting much more specialized. So if you wanted to do surveys, you probably use SurveyMonkey. They're really flipping good at surveys. They're not claiming to do anything else. The same thing is true with the cloud platforms. The people who built general case platforms are generally getting kind of pushed a little aside by more specialized offerings that are addressing narrower market segments better. How important is this issue of multi-tenant versus multi-instance? You obviously feel it's important. I mean, you guys are talking about it now. Let me put you in a hypothetical situation. You may or may not want to answer. Let's say you're a CIO, you're a big Oracle customer. Most of your CIOs here I guarantee are using Oracle in some way, shape, or form. Oracle's making a big push to the cloud. 12C, C for cloud, C for containers. I don't know, picked your poison. But Oracle's generally considered a pretty reliable company. Recovery is name of the game for them. They do a good job. Should I be concerned that if they're going in a multi-tenant direction, or is Oracle sort of an outlier in the cloud? You know, honestly, I'm not sure if they're an outlier, but I would say that if I were hired by Oracle to run their cloud, I would not do that, given their customer base. I do think there's a case where the early cloud companies, you sales forces, for example, were multi-tenant. They're multi-tenant because it was convenient. They're multi-tenant because that was their target audience. And so they were pitching, hey, look, the cloud. And that message ultimately got tangled up with their deployment architecture. So it's stuck in people's head that the cloud equals multi-tenant. And it really does it. SMB cloud, multi-tenant is probably exactly what you want to do. Departmentally focused, it's probably right. At the enterprise level, it's not the right design decision, though. Talk about what's new in the platform. Let's get into the platform. What's happening? Give us the update. Just the highlight reel real quick and then talk about what's exciting you about the next evolution of the platform. Sure. So a couple of different things. I'll talk a little about what we're doing for developers. Historically, I mentioned I talked about low-code developers, talked about no-code developers. There are also professionals. Now, I'm a professional developer. I did this for 20 years of my life. I lived in an IDE. I started writing code. I wrote C code. I wrote 370 assembler. I've done a lot of terrible, horrifying stuff back in my day. Terrible's probably the worst. School with bare feet and snow. Or natural. There you go. That's where to put it. It was really hard. I was being shot at, but no. The trick to that, though, is that if you were a professional and you wanted to use service now, the tools were not familiar. There was no IDE or single place you could go to see your whole app. So we built one. The Geneva release of the product actually has an in-browser IDE as code search. It has editing. It has code management. You see your whole app in one place. It's great. And actually, our teams use it to build itself. It's a little bit self-eating watermelon, but the team working on the IDE actually programs in the IDE. So they prefer that to programming in Eclipse, for example. We're biased. We like our IDE, but it's actually very valuable. So that's for the developer side. There's also a new developer program. You can go to developers.servicenow.com, join the program. You don't need to be a customer. Just have an email address. You can get a hold of a free instance. You can get access to technology. You can actually join the forums. As long as you use it, it's yours. It's really aimed at everybody. If you want to learn service now, go to the developer program, join it. There's no requirements other than a willingness to learn on your part. Technology-wise, though, I'll talk about something else. We live in a post-Edward Snowden world. And I don't really like Edward Snowden because he'd made my work harder. But one of the things he's done is make the concept of data sovereignty and data privacy a foreground concern for a lot of people, especially outside the US. People don't want to put data in the cloud if there's fear that a US-based vendor or US-based firms can potentially see it. We're set aside the US. If it's just private information, they don't want to put it in the cloud if anybody can see it. One of the ways to solve that, and we're addressing this, is to allow the data to get encrypted before it comes to us. So we're putting an encryption proxy inside the customer's network, along with its keys, and data will pass through the proxy. Certain fields get encrypted and we see only ciphertext. We literally can't read it. So encryption's your solution there? It is absolutely our solution. You can also add to the international law as you're going to create a replica, have a cloud-based system potentially, or can you store it in the US? It's stored in the US because the data is ciphertext. We literally can't read it. And there's side effects there that are actually kind of cool in that because we can't read it, you also can't use it in backend workflows. So you've got to design your app around the encryption. But that is a hard guarantee of it is we don't have the keys. It is not possible for ServiceNow to get your data back. And the government subpoena, you can't give them really anything either. No, you have to supreme the company in question who had the keys and up to their legal department as to what they wanted to do with it. Okay, so I'm going to ask you kind of as we wrap up here, a lot of great stuff. Containers are all the rage. I think Docker just got another $95 million. They've raised so much funding over the years. Containers but promises interoperability. I bring that up only as a way to tease out this notion of interoperability. How does that, how do you guys view that trend in the cloud? Is that something that's, containers have been around for a while? Sure. If you know programming, but Docker's got the traction and you're seeing security comes like Illumeo make it a lot of hype. I think there's two different parts to that. One is there definitely is a push to keep applications from messing each other up and impacting each other in bad ways either from a security standpoint or just from a architecture overload. And you see that on backend technologies. Docker is a good example of that. VMware's a little more mature technology doing something very similar. Zen, choose your virtualization layer. In the more application space where service now fits, we have the same problem in that we don't want a service now application to impact a different service now application. So we actually invested very heavily in Fuji something called scoping. It allows for applications to be managed individually, to be deployed individually and to be interact with each other only through defined APIs. And that means that you can actually deploy an application with a high degree of confidence. It's not going to impact any of the other for lack of a better word, innocent applications inside your system. It's a very big improvement. And it's one of the things that actually allowed us to do the service now store. How does open source evolution, if you will? You know, we always talk about this, you know, me being computer science degree back in the 80s, we lived in the same generations where open source was new. Second-class citizen now it's first-class citizen. Now you have beyond that now. It's proven, it's working. Is there a new business model as you're seeing kind of like pure, pure red hat? And then you're seeing, you know, open platforms like data platforms. So what's the next evolution of open source and how do you guys going to tap into that? And what's the most relevant thing for the folks to be looking at? Well, I think first of all, we're very big users of open source, especially in our backend. I mean, we're CentOS, we're a little bit of red hat, we're, you know, we're F5s, we've got Pixi, we've got Python, we've got Puppet, we've got lots of open source environment. And the product as well. We're huge fans. We think it really has brought a lot of really good technology out that's very accessible to the engineering community. So we use a lot of it. We even contribute back to some of them, as the case may be. I think if you look at business models, I'll be honest, I have not seen a lot of open source companies do really well in the environment. They've built a lot of great technology, and I think it's been very empowering for the developer community. But even red hat has not really, they're not huge. It's not a $20 billion company, the case may be. So I don't expect to see people flocking to the open source world to make money. I see people flocking to the open source world for the same reason engineers have always built cool stuff. It's that joy of creation, that power of building something. It's a real community of value, creation and contribution. It's that cycle of innovation. And it's not, I think, no one objects to money. I mean, that's why they call it money. But the open source world from what I've seen, it's not being driven by financials. It's being driven by engineers wanting to solve problems. It's kind of creativity for it. It's also a great way to play ball and get a job and show you what you're worth. It's like, you know, practice is like playing ball in the yard, sand a lot, baseball, then you go pro, right? So it's a way for recruiting and also to meet people. Absolutely. And we're actually, as I said, we're big users and we love a lot of technology. We're huge MySQL community users as well. Okay, Pat, we're going to get the hook here, but I want to give you the final word, the future. Give us your take of the preferred future, technology-wise and just next five years, 10 years. What's the world going to be like? I think five years out, it's going to look fairly similar to what it does today. You're definitely going to see a push to drive the information you need to you without you having to go and look for it. You're already seeing this, you know, Twitter pops when something happens, data comes to you. You don't have to go here and hit refresh periodically. That's going to drive itself into more and more parts of the world. Your iPhone dings when something comes up. That's going to seep out away from the phone, away from specialty platforms like Twitter and other applications. And you're going to get more and more used to seeing things come to you rather than you having to go out and look for information that's relevant. It's going to be kind of a service-oriented internet. It's going to kind of push stuff out to you. 10 years out, I suspect there'll be more dramatic changes. The big thing I actually see, this is a little bit of inside baseball, but operational architecture is getting much more standardized. So I do suspect that the amount of compute people can throw at problems is going to continue to go up astronomically. So right now, big data solutions are generally applicable to fairly narrow companies who can apply a lot of data to it. You know, like a Netflix can afford to optimize for recommendations for you. That compute's going to get cheaper and cheaper and more and more accessible. And you will see that sort of solution get applied to more and more specialized problems. So I think you're going to find that information is going to come to you and it's going to be more and more germane to you. Asynchronous, definitely. Absolutely. The value and the goodness of more and more cheap compute will create faster, faster personalization. Faster personalization and it'll be real time. There's no need for you to poll on it. It'll be asynchronous, it'll come to you and it'll be the information you need. Not near real time, real time. Self-driving cars don't do very well in near real time, do they? That's what we always say. Matt, thanks so much for sharing your time and insights here inside theCUBE. It's been a pleasure to get the insight from the early days to what's going on now. Appreciate it. This is theCUBE. We're live in Las Vegas for three days. For No 15, I'm John Furrier with Dave Vellante. We'll be right back with more theCUBE. Signal from the noise after this short break.