 This is Silicon Angles the Cube, our continuous production of AWS Summit. We're here live at the Moscone Center. My colleague and everyday co-host John Furrier is down at Stanford today at the Accel Partners, big data event. But we're here, we've been covering this wall to wall. Big data is obviously a hot topic. Machine data is one of the fastest growing and most underused by components of big data. And Jeff Sanjay Sarathi is here. He's the CMO of SumoLogic, a company that is in this field. Sanjay, welcome to the Cube. Thank you. So we're here at the AWS Summit. This is just unbelievable, the cloud is screaming, there's an old saying, big data gives the cloud something to do and machine data as I said is really growing super fast. So give us the update on SumoLogic where you guys are at and then we can get into the whole AWS angle. Sure, so SumoLogic is a company that was founded in 2010. And the entire premise is that the growth of machine data is the fastest growing component of big data. And this machine data can come from every part of your infrastructure, from applications, from your network infrastructure, from wireless devices. And the big challenge is not just collecting and aggregating and managing that data but actually generating insight from it. That's part of what we do. So how do you do that? So you've got all this data out there, you've got all this noise, this exhaust, I like to say. How do you guys make sense out of it and as we like to say, get signal from that noise? Sure, so a huge part of it is in our patent pending technological wall bridges which really solves one of the biggest problems around big data which is what do you do about data that you know nothing about? And so what we provide is a way to distill hundreds of thousands of log messages into a set of patterns. And those patterns are what provide that insight because you as somebody who's running that IT infrastructure actually knows what that message means. So those patterns tell you where to focus and give you the ability to essentially find that needle in the haystack without having to comb through every single part of the haystack first. So you're using math, statistics. Yeah, we have a lot of data science, machine learning is a huge part of our infrastructure. Every time you send data to us, we learn about what that data actually means. And when you interact with that data by telling us what's important and what's not important, that machine learning essentially takes that into account. And so the next time you do a query or an analysis on top of that machine data, we're actually providing you with the results that make sense for you and for your particular situation. So give me an example. So I've got my data, what do I do? I load it up into the cloud and I access your service. Take us through sort of a simple example. Sure, so you may have data in your data center, you could have data in the AWS cloud, you could have data off a network somewhere in your data center, and we'll essentially collect all that machine data typically in the form of walks and we'll then suck it into our cloud-based service and then we'll index it, we'll parse it, we'll allow you to immediately start creating searches of it because all of that data is collected in real time. We don't tell you to come back in an hour and then start searching. You can start searching as soon as that data gets indexed. Instinct gratification. Instinct gratification, better than an ice cream. So what happens is... Yet because it grows, it doesn't just go away. It doesn't just go away. And so what happens is as soon as you start getting that insight, you can then start performing analyses on it. You can start doing correlations across different parts of your infrastructure. You can perform statistical analysis on it. And then once you start doing that, you get a better handle of actually what's going on in the infrastructure. And so that gives you the ability to start baselining your infrastructure and say, this is really what the infrastructure is supposed to do. And then when anomalies occur, you instinctively know what those anomalies mean because they're 10% above or 10% below that baseline that you've set for your infrastructure. So what are people doing with your product? We talk about some of the applications and use cases. Sure, so for example, Netflix is a client of ours and they use us to manage their entire internal IT infrastructure. One example, they collect logs from their virtualization infrastructure, which is VMware and they troubleshoot it, monitor it, correlate it with other parts of what's going on in their infrastructure. We have a customer called Apigee, which is the API company. And they use us for a couple of different purposes. They use us for compliance purposes so that, because they need a centralized way to collect, manage and store all their logs to compliance purposes, and they use us as a single system of record for them. They also use us to essentially monitor the errors as part of the application development process. Again, so the use cases range from application monitoring, they range from application management, operations management to compliance and security. Okay, and then, how do you charge for your, is it a service, is it a product, a combination? Yeah, it's a service, it's a cloud-based service, so we charge on a subscription basis, and it's based on really two factors, how much data you want us to ingest and analyze, and how long you want us to store the data. Okay, so you charge by the terabyte of ingestion? Yeah, by the gigabyte, by the terabyte of data. And we're very flexible, so it's not based on how much data you're ingesting on that behalf, we normalize it over the course of a month. So if you need to elastically scale to handle one day's worth of bursting data, we can do that, and it won't affect necessarily your payment without spending a cost. So what's your relationship with Amazon, how do you leverage Amazon web services? Sure, we're built entirely on the Amazon cloud. When we started the service in 2010, we realized very quickly that Amazon essentially turns the data center into an API, and we want to essentially hire developers who can write to that API, we don't want to hire data center experts. And so we run on top of EC2, we run essentially in S3, because we store all our customer logs as well as our own production logs in S3, and we use DynamoDB for all our metadata storage. Okay, so Dynamo is your key value store and allows you to scale and service. Okay, good. So talk about this event. What have you guys got going at this event? Maybe some of the customers that you've talked to, what they're saying, what's the buzz line? The buzz has been great. I mean, we were at the New York event as well, and so Anna, the first reinvent event and back in November. What we found at these events, it's both a great learning experience for our team from our engineer perspective for some of the technical sessions, and it's a great place to do business with both existing customers who are here as well as prospects who we've been able to meet. So can you, we talked earlier about the sort of machine data that's underutilized, growing fast. Can you quantify that? Yeah, there's more data that's going to be generated, more machine data generated in 10 seconds in 2013 than was generated the entire year, a decade ago in 2003. So there's more data being generated across every part of your infrastructure, and the big challenge early is how do you create insight out of that and what do you do with it? And most companies are talking with a large, Midwestern consumer goods company, and they said, we have a farm of 3,000 web servers, and we have no idea what data is being generated from that, and we have customers buying product off that web server farm, and they said, can you help us? And we said, yes, I think we can. So who do you sell to? Who's kind of like your perfect profile of buyer that you're interacting with? So our target audience typically is the IT audience. We sell to the DevOps community, we sell to VPs of application management, we sell to VPs of operation. Ultimately, they report typically up into the CIL, and this is not just a technical purchase decision, it's a business purchase decision, because a lot of the insights that are being generated by those machine data affect how customers are buying, what customer service is being impacted potentially by outages in your infrastructure, how you're dealing with porters, and so a lot of the machine data actually impact day-to-day business decisions, which is why oftentimes we start with IT and the decisions made with IT, but it's through the business use case of mine. Talk about that business use case, that economic model, you know, the ROI. I was curious, what's usually the first business use case to get them to start paying attention to logs? Yeah, a lot of times it's supporting an SLA that's in the place. So if a company has a 3.9 or a 4.9 SLA, then you need to support that because that actually means dollars out of your pocket if you're not there. Then companies use Sumo Logic to meet their SLA to understand where they may be, where their applications may be running more slowly than expected, where they may have downtime in the infrastructure they need to solve. There's a great example of a customer that was in our conference from one day and we were showing them the power of logics, the ability to create patterns from this data. And they had five or six of their IT staff there and we were showing them these patterns. And all of a sudden the conversation stopped, they started talking amongst themselves and three people just left the conference room and said, excuse me, we now know what's wrong, we need to go solve it. There was an SLA issue and we needed to figure out what was going on with the data. Was that hypothesis based when you started playing with the data? Or is it still in the surface? It just surfaces it. We don't necessarily know what the hypothesis is. We just know what the patterns are. But that's why the combination of humans plus machine learning is very, very powerful. You as a human know what those patterns actually mean. And so you can then delve into those patterns to figure out root cause analysis and then figure out what's actually going on. Yeah, interesting. So Shift Gears a little bit and talk about you're starting a business in 2010 that's super data intensive. And the fact that you decided to use AWS as your infrastructure. Yeah, for us it was absolutely a no brainer. We knew right off the bat we didn't want to build our own data center. We knew that we wanted to get to market quickly. And with all those considerations in mind we said it absolutely makes sense for us to build this in the cloud. And at the time it made absolute sense to look at AWS and say, hey, do they provide us with the necessary infrastructure to make that happen? And the more we examined it it just became a no brainer, we just said. It's in the cloud. It's going to be a service. AWS provides us with everything. I mean economics have to make sense first and foremost. We're building a business on it. Yeah, and the economics the economics made sense both when we started off and as we've grown over the past three years considerably at scale they make even more sense. And that's been from a TCO perspective it makes sense for our customers as well because they've been able to offload the cost of managing servers and storage for all the log and machine data and people to handle all that. And so when we go into prospects and show them the TCO argument of a cloud-based service versus an on-premise-based environment it's a very easy decision to make. Wow. Sanjay, talk about how you differentiate from the competition. There's a lot of buzz now in the marketplace. You've got obviously you've got Splunked at an IPO. You've got guys like HP and IBM saying, oh, we can do that too. You guys 2010, so you're bringing a 2010 on perspective. How are you different than the other players out there? Yeah, I think it boils down to a couple of things. One is in the whole analytics sphere, the whole notion of being able to create patterns from the amount of machine data that you have and those patterns and the insight you generate from those patterns are a lot of times a wild factor when we go into clients accounts because they've never seen those patterns before. The second really is the elastic scalability of our service and certainly sitting on top of Amazon helps us. But the ability to immediately support a customer that needs to burst 5x, 10x the amount of data that they did potentially in the prior day and then to skim up and down is something that's a huge value add to our customers so that they know that they can handle customer surges whenever they might occur. So you're all in on cloud. Anil Bhushri is on the board obviously. He knows a little bit about Cloud Workday. Co-CEO of Workday is the hottest IPO out there and excellent, great story. Really appreciate you sharing it with us and thanks for coming on theCUBE. Good luck and we'll probably see you at re-invent. Absolutely, thanks a lot for having me. All right, it's a pleasure to meet you. Thank you. All right, everybody, keep right there. Jeff Frick and I will be back to wrap up the day from AWS Summit. We're here at the Moscone. This is theCUBE. We're live. We're here all day wall-to-wall coverage. We've got our colleagues down in Stanford, John Furrier's down there. Check out SiliconANGLE2 for that channel. We'll be right back with our next guest and our wrap up right after this.