 Live from Washington, D.C., it's theCUBE. Covering.conf 2017, brought to you by Splunk. Welcome back to the nation's capital, everybody. This is theCUBE, the leader in live tech coverage, and we're here at .conf 2017 Splunk's customer event. This is the seventh year that we're covering .conf with theCUBE here in the nation's capital in the district, I'm Dave Vellante with George Gilbert. For the wrap of day one, we'll be here for two days. George, good day overall. Splunk, the Splunk ecosystem continues to grow. Splunk evolves as a company. We're talking about a company, we didn't really have time this morning to run this down, but it's about a $1.2 billion company, growing at around 30% a year. It's got a $10 billion market cap, thanks in some part to the Symantec CEO, opined that, hey, Splunk might be a good acquisition target, and the stock shot up there for a little bit. 15,000 customers, they got a billion dollars in cash, zero debt. So nice balance sheet, good growth, small but meaningful, positive free cash flow. So from a financial perspective, Splunk's looking pretty good right now, new CEO. They had some bumps in the road in the past, some kind of guidance issues, but all seems to be pretty good right now. From your financial analysts, put your financial analysts hat on for a second, how's the company look to you? I actually think that numbers look better than the sort of high level optics, because it's mostly a subscription revenue, and so you're, rather than get, say, $100 up front from a perpetual license, they're getting, say, $20 to $25 over a period of many years. So that actually depresses your operating margins. Sure. And so their revenue, their sort of revenue impact and their profitability is better than it looks. Am I mistaken? I thought the vast majority of their revenue was still perpetual license, right? I think they've been converting to where you pay on the throughput, how much data you ingest per day. And I think that that's, you don't pay for it all up front. So they're migrating to a rateable model, which is oftentimes the crushes companies, but they seem to be managing through that. So anyway, that's sort of one thing that I wanted to talk about a little bit. Some of the themes that you and I talked about this morning, there were six that you and I kind of laid out. The expansion of the total available market, really from a monitoring log data into sort of more of an application platform. Part of that is the shift from SIM, from a security standpoint, into more analytic oriented activities. The second one was the whole cloud and hybrid cloud play. Another theme we looked at was admin and dev complexity and Splunk's recipe for simplifying that. Machine learning, where does that fit in? Obviously with some of their ITOM stuff, they're trying to be more proactive and anticipatory. Breath or depth, meaning do they go deep within a sort of an application silo or a use case or are they sort of more broadly based platform? And then the last one, number six, sort of IoT and edge processing. George, that's not something that we were able to spend much time on this morning or any time. So I'd like to start there. Everybody talks about IoT. We all know that at least in concept, all this data is going to be generated. A lot of it is stateless. We talked about that on the Wikibon research meeting a couple of weeks ago with serverless. Question, where does Splunk fit in IoT? If the strategy is to sort of send it all back to the cloud, is that a viable approach? And is that their strategy? It's not their strategy. It's what their architecture allows today, but they know that doesn't work because in a world of sort of industrial assets and sort of consumer devices, you're producing so many more devices per year and so many more data elements per device, per time period, that the amount of data is exploding exponentially, you cannot for latency and bandwidth reasons send that all to the cloud to get an answer and then send it back. So part of what's happening and part of what Splunk is building is the ability to capture at data, perform low latency analytics, drive an answer to a local device and then what they do is what other IoT platforms do, send up the interesting data, the stuff that doesn't fit, the stuff that you want to make sense out of where you have to rethink your model, your predictive model and then that sort of research and refinement happens in the cloud and when you think you have a good new model, you push it back out to the edge. This is again all theoretical, they haven't talked about it yet other than directionally, but it's worth saying as our distinguished CTO reminds us that something, David Floyer, 90, 95% perhaps of the data and the analytics will happen, really the data processing will happen at the edge. More interesting though is the division of labor up in the cloud, it's not just retraining a model but we'll have very rich simulations. So rather than just saying training a self-driving car to in the snow to avoid sunlight that obscures its view of the hazards in the road, you actually might have a simulation where you go through a whole bunch of different essentially edge conditions. So the models get very, very rich and then those get pushed down to the edge for local processing. And then learning is iterative. Yes, yes. And that continues. And okay, so that's cool. That sort of leads to the discussion of cloud and hybrid cloud. We heard even from AWS that much of the processing and analysis can occur on-prem in their model. It's not something that just has to get done in the AWS cloud. Interesting to hear AWS acknowledge that. Whereas five, six years ago, sort of their dogma was everything goes into the cloud. So they're learning and evolving along with their partners. But what about Splunk's cloud play? Years ago they announced a cloud offering. We talked earlier about much more of their revenue coming from ratable models. I think 50% of their new business is cloud only which makes sense. A lot of data analysis is going on in the cloud. What's your sense of their cloud strategy? Is it working? Are you sanguine toward their approach? So we've had, since the dawn of the Pleistocene era in computing, we've had multiple platforms. And there has always been a desire to have a common development and runtime environment across different platforms so that developers are not locked in or so that they can have a common platform for building apps across platforms and for running them. The same like so that you had part of Cisco's success and Oracle's success was you had the same admin experience no matter what you were running on. So Linux obviously addressed what Unix never could. It was the promise of Unix. Obviously a lot of Microsoft's ascendancy was given that binary compatibility with Windows. Okay, so will we achieve that with cloud? It looks like we're further away from that than ever. There's choices here where with Splunk, they will have this self-contained environment that can run on many platforms. They'll run on-prem, they'll have some subset that runs on the edge, they'll have something that runs compatibly on Azure and Amazon and Google. But once they're on the cloud, there are these really powerful centrifugal forces that are pulling apart the compatibility of that singular platform because you'll have very specialized services. For instance, if you're doing IoT with Amazon, you have the Kinesis Firehose service that's pumping data into Splunk or into S3 where other services might be operating on it. Whereas with Azure, you might have different edge services pumping data into, could be Splunk, could be Splunk plus other services. For instance, Splunk doesn't have a really strong scale out SQL database where you might want to do some advanced analytics as part of your predictions. Okay, so I could leverage DynamoDB as an example and something like that. Or Redshift on Amazon or Snowflake as a cross-platform, that sort of thing. Okay, good. You're here tomorrow, yes, at least in the morning? Yeah. Okay, homework assignment tonight. So were you participating in the analyst event today? Okay, so you got some other insights. So bring all the non-NDA stuff. Tonight, like I say, a homework assignment. Try to distill that down. We'd love to have you back if you have the time at the open tomorrow. If I have the time to save. I flew across the country to sit next to you and like. That's awesome. Great, all right, good. So boil it down for us. Tomorrow, why don't you come on and take us through kind of what you learned yesterday, maybe some of the product announcements and give us your, the George Gilbert, kind of Wikibon view of the future for Splunk and this industry, okay? Okay. All right, great. Thank you, George, for helping me wrap. That is a wrap, day one today. This is theCUBE. We're live all day tomorrow. Watch the replays at siliconangle.tv. Check out siliconangle.com for all the news. Check out wikibon.com for all the research and go to Twitter. The hashtag of this event is Splunk Conf 17. And also check out hashtag CUBE Gems and you'll see the snippets of today's show. This is theCUBE, the leader in live tech coverage. We're out day one from the district. See you tomorrow.