 Hi, from San Francisco, it's The Cube. Hi, welcome back. Jeff Frick here on the ground at the Cassandra Summit 2014 at the Western St. Francis in San Francisco, California. I think this is the third or fourth year. I can't remember. Is it the fourth year they said they've had Cassandra Summit? New venue, new energy, a lot of excitement. I'm joined here in this next segment by Robbie Strickland from the Weather Channel. So Robbie, thanks for stopping by. Thank you very much. Thanks for having me. So you walked up. You got the speaker badge on, so you're obviously an important person, and as you know, we like to get the smartest people we can find and ask them great questions. So what are you going to be talking about later today? Well, if you're looking for the smartest people, I'll have to go find them and bring them back. I'm talking about CQL under the hood, which is basically the new, latest, greatest way to talk to the Cassandra database. We used to talk with an old protocol called thrift, and a lot of the old guys, they're having a hard time moving from one to the other, and a lot of the new guys, they're seeing something that looks familiar to some of the stuff they've seen in the past with their relational databases, and I'm trying to help them understand what's actually happening under the hood so that the old guys can make the transition effectively, and that the new guys cannot make some of the pitfalls that I've seen out there. And is this new technology, new features in a recent release, or you're just still trying to get everybody up to speed? I would say it's more still trying to get everybody up to speed. It's been out for a year and a half, two years, but it's been a slow progress. And why? Just people aren't educated? Or they're sticking with the tools that they know and love, or is it just new? Why the slow progress? I think the slow progress, because in the new, the no-SQL space, people are used to doing things in a little bit different ways, so the veterans of Cassandra are not, they don't like the fact that they've got a new syntax that looks a lot like the old stuff that they used to use in the relational database world, and there are a lot of reasons for that, for the trepidation there. And I was actually one of those people, but I've since converted, and one of the other things that I've seen is that new people, they see something that looks a lot like what they did with Oracle, or MySQL, or something like that, and that can also be problematic because they can fall into some of the traps that, where, hey, this looks like the old stuff, but I need to do it a little bit differently, and I'm trying to help people understand what they need to do. Yeah, it is one of these weird things. When things aren't really different, and they're a little bit similar to what you're familiar with, it's almost harder to adopt them, and it's just completely radically different, and now you just know you're doing something different. That's right, it certainly can be that way. So, the Weather Channel, you guys obviously process a ton of data, we're getting better at forecasting, hopefully, with each and every day. I wonder if you can share some of the statistics behind what you guys are working with in terms of the scale of the data, the distributed data, and why you went with something like Cassandra. Well, we actually processed about 10 billion transactions a day, which it's not Twitter scale, but it's still, it puts us in a category that is somewhat rarefied territory, and so in order to do that, we need databases that scale. We also have clients that consume our data all over the world. If you use things like iOS 8, or Google Now, or Yahoo! Weather, all that data comes from us. A lot of people don't know that. That's actually the biggest part of our business, is serving data to all these other third parties. We're also the number one on web with weather.com. We have the number one mobile apps on all platforms, and so, and obviously, we have a TV station, which is the thing most people even think about. But that's actually quite small when it comes to the amount of data that we process. So we're doing all that. And now we just started moving to a forecast on demand. So when you ask for a forecast for your latitude longitude using your mobile app, you get a forecast generated for you in that location with very fine granularity on the spot. And so that's pretty exciting for us. And so we actually run Cassandra data centers all over the world. So we're a multi-data center deployment. So talk a little bit about how the changing data types and data available to you guys has changed your business. Because clearly there was probably a limited number of sources you had early on. But now there's more and more things out there. Now we're all packing our own little mobile weather stations, if you will. I don't know if you're tapping into that. How is that kind of changing your business, both in terms of what you can offer to your customers, but then also on the back end in terms of the technical challenges? Well, one of the really interesting things that we're doing actually on that front that is powered by Cassandra is we have a feature called social weather, where you're able to tell us when you get your, in our newer Android and iOS applications, you're able to tell us, hey, I don't agree with you. You told me, you're telling me that it's raining right now, and it's actually sunny, or vice versa, or something like that. And so we're able to take that input from everyone and sort of put it into our models, which is pretty exciting. Talk about everyone's favorite weather challenge, right? The difference between forecasting in the short term and the long term. Is that getting easier? Is it a different algorithm? I mean, how do you guys wrestle with that? Because we all want to know what the weather is going to be like tomorrow. But then there's a lot of bigger, longer term trends that people are keeping track of and are relevant. So does it all fit into the same models, the two different models? How do you kind of look at the short term versus the long term? Well, I am not a meteorologist. So I can't tell you a lot about forecasting. But I can say that our model, we measure accuracy. That's our number one metric of success. And I can tell you that we're getting more and more accurate over time. And if you compare us to all of the other people, our accuracy is far and away better than all the other guys. So a little commercial for the Weather Channel. A little commercial for the Weather Channel. And then I didn't ask about microclimates between the Mission and downtown and Sunset District. If you're from San Francisco, that could be a... San Francisco is a challenging climate in no uncertain terms. Well, I mean, it begs the question, will you at some point in time allow me just to send you the meteorological information that my phone can pick up? Well, we already do through, because Weather Underground is actually part of the Weather Channel as well. A lot of people don't realize that. And Weather Underground has thousands and thousands of personal weather stations all over the world. And so that's actually one of our sources of weather data. So we've got them all over San Francisco. So we can see that on Telegraph Hill, it looks like one thing, and in the Mission District, it looks like something else. And so actually, that's one of the reasons why our forecast accuracy and our current data is better than other people's. Great. Well, Robbie, thanks for stopping by. Good luck teaching the old dogs new tricks. The new dogs old tricks. Good luck on your presentation. I'm Jeff Frick. We're on the ground at Cassandra Summit 2014. The West in St. Francis in San Francisco, California. Thanks for watching. We'll be right back.