 Hi, Jeff Rick here. We are on the ground at Spark Summit 2015 in San Francisco. As you've heard, Spark is like the hottest thing going in big data. We were at Hadoop Summit last week, and everybody's talking about Spark, so we want to come up and check it out, see what's going on. We're really excited to be joining this next time, my many-time Cube alum, Gary Ornstein, who was most recently at Fusion I.O., left a little while ago after the acquisition. We've been waiting for him to surface, and sure enough, where do we find him here at Spark Summit? So, Gary, good to see you. Great to see you too, Jeff. So tell us, where are you? What are you up to? So I'm working at MemSQL. We are a company that makes a real-time database for transactions and analytics. And we're here at the Spark Summit because we also have a MemSQL Spark connector to help take Spark results, put them into an operational database, and put them to use. So, Gary, you've been in the industry a long time. Why is everybody so jacked up and excited about Spark? Well, I think what Spark has done is take an enormous amount of features and functions in terms of programmatic processing framework and make it simple and easy for people to operate from that single Spark console. And so this wide range of functionality that can be easily accessible by data scientists. So I think that's driving a tremendous amount of interest, the performance, the capabilities, and it also fits with what's going on with our world. We live in a real-time world, nobody wants to wait, and a memory-optimized architecture like that that's part of Spark or part of MemSQL for that matter is really appropriate for today's workloads. And then talk about how the workload world has to change with this whole concept of a data pipeline, right? This constantly streaming data versus grab it, hold it, and do something with it. Exactly. We've been seeing a lot of pickup with customers, something we call the real-time trinity of data pipelines, which is a combination of Kafka for the message queue, a combination of Spark for real-time transformation, ETL, data enrichment. And then Spark itself doesn't have a persistence layer, so you have to put that data somewhere, and if you put it in a database like MemSQL, you can then build a rich SQL-based application right out of the box. It can be something that's done with custom dashboards or using an off-the-shelf application as simple as Tableau. Awesome. So you said you've got some customers that are actually out there in the wild using this stuff. I wonder if you can give some examples of what they're doing. Yeah, one of the greatest examples is Pinterest, who built a real-time data pipeline using Kafka, Spark, and MemSQL. They wanted to track what was happening with repins, which is an indication of viral content. So they tracked all the repin information, pushed it through Kafka, did enrichment inside Spark, pushed it into MemSQL, and then they were able to build a rich visualization app just using SQL queries to see what was happening with repins in real-time, giving their data science team real insight to what was happening at that very moment. And the end-to-end latency on a data pipeline like that can be less than one second. So when you're running... Less than one second from the time my wife Joy pinned something, because she's a big Pinterest gal. Yes. This shows up on there. From Joy doing a repin to that showing up in the query. So we like to say, this is not about what was. This is about what is. What is, right? And tools like Spark being a memory-optimized, distributed system, MemSQL being a memory-optimized system, Kafka, for that matter, being memory-optimized, it really is taking us to a whole new world. And you're all about the flash, right? So we know you like stuff to move fast. So kind of last question, right? We always talk about people, process, and technology. The technology's here. What has to change on the customer side in terms of what they do, how they do things, when suddenly they've got all this real-time data? Are they ready for it? Do they know what to do with it? Well, I think a lot of people are still capturing data and just putting that into HDFS. And that's been one of the challenges of how do we actually make use of that data? So I think what we're seeing is when you put some of this incoming stream into a structured database such as MemSQL, everything's just a SQL query away. It also takes a lot of time to showcase what this can do inside an organization. And so what we encourage folks to do, we have a new community edition of MemSQL that's free, unlimited scale, unlimited capacity, is build something simple. Build a real-time data pipeline. Don't make it too complicated. Stand it up. Show it to your colleagues inside the office and inside of work and see what you can do with these new tools and technologies. Awesome. Well, Gary, thanks for stopping by. Yeah. It's always good to see. Keep an eye on Gary. Keep an eye on MemSQL. It doesn't go places unless there's some good action going on. So I'm Jeff Frick. We're at the Spark Summit 2015. Joined here by Gary Ornstein. Thanks for watching. Take care.