 Live from San Francisco, it's theCUBE, covering Google Cloud Next 19, brought to you by Google Cloud and its ecosystem partners. Hey, welcome back everyone live here in San Francisco, California. It's theCUBE's coverage of Google Cloud Next 2019. It's our third day of three days of wall-to-wall coverage. I'm John Furrier, my co-student. Dave Vellante's out and around on the floor getting stories, getting the scoops. Of course, we're here with Sudhir Hasbay who's the director of product management at Google Cloud. Sudhir, great to see you again. Thanks for coming back on. Back on last year, obviously big query was a big product that we love. We've talked to Dipty many times about database. We geek out on the databases, but it's not just about the databases. We talked about this yesterday all morning on our kickoff. There is going to be database explosion everywhere. Okay, there's no one database anymore. It's a lot of databases. So that means data in whatever database format, document, relational, and structured, whatever you want to call it is going to be coming into analytical tools. This is really important. It's also complex. It needs to be made easier. You guys have made some serious announcements. So let's get to the hard news. What's the big news from your group around BigQuery ML, AutoML, some of the news? Share the news. Perfect, I think not just databases are growing, but also applications. There's an explosion of different applications. Every organization is using hundreds of them, Salesforce, to work a day, to so many of them. And so having a centralized place where you can bring all this data together, analyze it, and make decisions is critical. So in that realm, to break the data silos, we have announced a few important things at the event. One is cloud data fusion, making it easy for customers to bring in data from different sources on premises in cloud so that you can go ahead and, as you're bringing the data in transform and visually just go ahead and move the data into BigQuery for analysis. The whole idea is to go ahead and have drag and drop code-free environment for customers to easily bring data in. So we have a lot of customers just bringing in all the data from their on-premises systems, Oracle, MySQL, whatever, and then moving that into BigQuery as they analyze. So that's one big thing. Super excited about it. Lot of good traction, lot of good feedback from our customers at the event. The second thing is BigQuery, which is our cloud-scale data warehouse. We have customers from few terabytes to hundreds of petabytes with it. We also have an in-line experience for customers, like a data analyst who wants to analyze data, let's say from Salesforce, Workday, or some other tools like that. If you want to do that, we have made 100 plus connectors to all these different SaaS applications available through our partners, like FiveTrans, Supermetrics, in BigQuery. Five, four to five clicks. Out of the box. Out of the box, four to five clicks. You will be able to go. In the cloud, but not a box, but I guess that's the phrase. But it's important. Connectors, the integration points are critical table stakes now. You guys are making that a table stake, not an add-on service to pay. You just basically go in and do five clicks, you can get the data in, you can use one of the partner connectors for making all the decisions and all. So that's there. And we also announced migration service to migrate from Teradata, Redshift, those things. So just making it easy to get data into GCP so that you can unlock the value of the data is the first thing. This has become the big story here from the CUBE standpoint. And Stu and I have been talking about data all week. Data migration has been a pain in the butt. And it's the critical linchpin that some say could be the tell sign of how well Google Cloud will do in the enterprise because it's not an easy solution. It's not just, oh, just move stuff over. Enterprises have unique requirements. There's all kinds of governance, all kinds of weird deal, things going on. So how are you guys making it easy? I guess that's the question. How are you going to make migrating in good for the enterprise? Yeah, I think the one thing I'll tell you just before that, I had a customer tell me one time, you have the best highways, but your on-ramps to the highways are a challenge. Can you fix that? And I'm like, yeah, absolutely. It's a great analogy. Yeah, it's great. And so last year or so we have been focused on making the migration really easy for customers. We know a lot of customers want to move to cloud and as they move to cloud, we want to make sure that it's easy drag drop, click and go for migration. So we're making that whole- So figure the on-ramps, basically get the data in. Yeah. What's been the big challenge? What's the big learnings? What's the big accomplishment? I think the biggest thing has been in past people had to write a lot of code to go ahead and do these kind of activities. Now it is becoming click and go, make it really code-free environment for customers, make it highly reliable and all. So that's one area. But that's just the first part of the process, right? What customers want is not just to get data into cloud, into BigQuery, they want to go ahead and get a lot of value out of it. And within that context, what we have done is, we've made some announcements in that area. One big thing is the BI Engine. So BigQuery BI Engine, it's basically an acceleration on top of BigQuery. You get like sub-second latency response times for interactive dashboarding, interactive reporting. So that's there. But with that, what we have also announced is connected sheets. So connected sheets is basically going to give you spreadsheet experience on top of BigQuery data sets. You can analyze up to 10 billion rows of data in BigQuery directly with drag drop. You can do pivot tables. You can do visualizations. Customers love spreadsheets in general. Yeah, Sudhir, I'm glad you brought it up. We run a lot of our business on sheets. We have so many of the pieces there. And right, if those are the highways we're using our data, what's the first steps? How do we start? What are some of the big use cases that you see with that? So I think AirAsia is a good example. So AirAsia has a lot of their users, operational users, you needed to have access to data. And so they basically, first challenge was, they needed to really have a sub-second latency so that they can actually do interactive access to the data and all. So BI Engine has helped with that. They use Data Studio on top of BigQuery to go ahead and make it accessible. BI Engine will work with all the other partner tooling too. But on the other side, they also needed to have spreadsheet-like, really complex analysis of the business so that they can improve operation. Last year we announced they have saved almost five to 10% on operational costs. And in an airline, that's pretty massive. So basically, they were able to go ahead and use our connected sheets experience. They've been an early alpha customer to go ahead and use it, to go ahead and analyze the business, optimize it and all. So that's what customers are able to do with connected sheets. Take massive amounts of data of the business and analyze it and make better decisions. How do we use that? So if we're a customer, pretend we want to be a customer, we have so many tweets and data points from our media, I think 15 million people are in our kind of Twitter network that we've indexed over the years. I try to download on a CSV, it's horrible. So we use sheets, but also they've had limitations on that client. So do we just go to BigQuery? How would we work it? You can use Cloud Data Fusion with few clicks, move data into BigQuery. Once you now have it in BigQuery, in Sheets, you will have an option from data connectors to BigQuery. And once you go there, if you're in extended alpha, if you should get into extended alpha, and then once you click on that, it will allow you to pick any table in BigQuery. And once you link the sheet to BigQuery table, it's literally the spreadsheet is the front end to the whole BigQuery. So when you're doing pivot tables, when you're saying, hey, aggregate by this and all, it actually is internally calling BigQuery to do those activities. So you remove the barrier of doing something in the presentation layer and move that to the engine that actually can do the large-scale processing. Is this shipping now you mentioned is an extended beta? What's the product? It's an extended alpha for connected sheets. So it's like we're working with few customers early on to go ahead and make sure. So you guys are doing lighthouse accounts, classic? Lighthouse accounts, yeah. Classic early. If customers are already a G Suite customer, we would love to get more. Is there a criteria on the connected sheets, alpha, extended beta, alpha? Extended alpha. What's the criteria? I think nothing. If customers are ready to go ahead and give us feedback, that's what we care about. So you want to start with like 20, 25 customers and then expand it over this year and expand it, maybe make it available by then. So people watching, let us know what's the URL, where do they go get it? Just tweet to me and then I can go connect with the right folks. So, Sudhir, one of the other announcements I saw this week, I'm curious how it connects into your pieces is a lot of the open source databases. And Google offering those as a service. Maybe you can expand as to, because we know, as John said in the open, the proliferation of databases is only going to increase. I think open source, we announced a lot of partnerships on the databases. Customers need different types of operational databases. And this is a great opportunity for us to partner with some of our partners and providing that. And it's not just databases, we also announced partnership with Confluent. I've been working with the Confluent team for last one plus year, working on the relationship, making sure our customers have an op. I believe customers should always have choice. And we have our native service with Cloud Pub-Sub. A lot of customers like Kafka, they're familiar with Kafka. So with our relationship with Confluent and what we announced now, customers will get native experience with Kafka on GCP. I'm looking forward to that, making sure our customers are happy. And especially in the streaming analytics space, where you can get real-time streams of data, you want to do directly analytics on top of it, that is a really high value add for us, so that's great. And so, what I'm looking forward to is customers being able to go ahead and use all of these open source databases, as well as messaging systems to go ahead and do newer scenarios with us. Okay, so you got BigQuery ML, was announced in GA. BigQuery also has auto support, auto ML for tables. What does that mean? What's auto ML for tables? So we announced BigQuery ML at cube last and next in beta. So we are going GA with that. BigQuery ML is basically a SQL interface to creating machine learning models at scale. So if you have all your data in BigQuery, you can write two lines of SQL and go ahead and create a model to, with let's say clustering, we announced clustering now, we announced matrix factorization. One great example I will give you is booking.com. So booking.com, one of the largest travel portals in the world, they have a challenge where all the hotel rooms have different kinds of criteria which says, hey, I have a TV, I have all the different things available and their problem was data quality. There was a lot of challenges with the quality of data they were getting. They were able to use clustering algorithm in SQL, in BigQuery, so that they could say, hey, what are the anomalies in this data sets and identify, there were hotel rooms that would say, I have a satellite TV, but no TV available. So those data cleansing stuff, they were easily able to do with a data analyst with SQL experience, so that's that. That's a great example of automation. Humans would have to come in and clean the data that manually and or write scripts. So that's there, but on the other side, we also have amazing technology in AutoML. So we had AutoML vision, AutoML available for customers to use on different technologies, but we realized a lot of problems in enterprise customers are structured data problems. So I have data in BigQuery, I want to be able to go ahead and use the same technology like neural networks to go create models on top of that data. So with AutoML tables, what we are enabling is, customers can literally go in AutoML table portal, say, here is a BigQuery table, I want to be able to go ahead and create a model on. Here is the column that I want to predict from based on that data and just click a button, we'll create an automated the best model possible. You'll get a really high accuracy with it, and then you will be able to go ahead and do predictions through an API or you can do bulk predictions out and store it back into BigQuery and all. So that's the whole thing, when making machine learning accessible to everyone in the organization, that's our goal. And with that, with BigQuery and AutoML. And a better end product, too. Yeah, exactly. It should be inbuilt into the product. So we know you've got a lot of great tech, but you also talked to a lot of customers. I wonder if you might have an example so to really highlight the updates that you have. I think booking.com is a good example. 20th Century Fox last year shared their experience of how they could do segmentation of customers and target customers based on their past movies that they had watched. And now they could go ahead and predict. We have customers like News UK, they are doing subscription prediction, like which customers are more likely to subscribe to their newspapers, which ones are trying, may churn out. So those are key examples of how machine learning is helping customers, like basically to go ahead and target better customers and make better decisions. So do you talk about the ecosystem? Because one of the things we were riffing on yesterday, and I was giving a monologue that Dave about, we had a little argument. But I was saying that the old way was a lot of people are seeing an opportunity to make more margin as a system integrator or global SI, for instance. So if you're in the ecosystem dealing with Google, there's a margin opportunity because you guys lower the cost and increase the capability on the analytics side, mentioned streaming analytics. So there's a business model money making opportunity for partners that have to be kind of figured out. What's the equation there? Can you share that because there's actually an opportunity because if you don't have to spend a lot of time analyzing the content from the data, talk about the money making opportunity. There's a huge opportunity with global system integrators to come in and help our customers. I think the big challenge is more than the margin, there's a lot of value in data that customers can get out of. There's a lot of interesting insights, lot of good decision making they can do. And a lot of customers do need help in ramping up and making sure they can get value out of that. And it's a great opportunity for our global SI partners and I've been meeting a lot of them at the show to come in and help organizations accelerate the whole process of getting insights from their data, making better decisions, do more machine learning, leverage all of that. And I think there's a huge opportunity for them to come in, help, accelerate organization. What are some of the low hanging fruit opportunities? I'll see that on ramping or the data ingestion is one. What's the low hanging fruit? Yeah, so I think low hanging is just moving migration. Earlier we said break the data silos, get the data into GCP. There's a huge opportunity for customers to be, like, you know, get a lot of value by that. Migration is a huge opportunity. Lot of customers want to move to cloud, then they don't want to invest more and more in infrastructure on-prem so that they can leverage the benefits of cloud. And I think helping customers migrate, migrations is going to be a huge opportunity. We actually announced a migration program, also like a week back or so, where we will give training credits to our customers. We will fund some of the initial, we'll put initial investment in migration activities with our SI partners and all. And so that should help there. So I think that's one area. And the second area I would say is, once the data is in the platform, getting value out of it with BigQuery ML, AutoML, how do you help customers with that? I think that will be a huge opportunity to. So you feel good, Sudhir, but you're going to build an ecosystem. Yeah. You feel good about that? Yeah. We feel very strongly about our technology partners, which are folks like Looker, like Tableau, like Talent, Confluent, Trifacta for Dataprep. All of those, that partner ecosystem is there, great. And also the SI partner ecosystem for delivery so that we can provide great service to our customers will be important. Gets some good logos on that slide. I got to say Trifacta and all the other ones, we're pretty good, Splunk, et cetera. Well, okay, so what's the top story for you in the show here? Besides, you grew out on the data side for your area. What's the top story? And then generally, in your opinion, what's the most important story here at Google Cloud Next? I think two things. In general, the biggest news, I think, is the open source partnerships that we have announced. I'm looking forward to that. It's a great thing. It's a good thing, both for the organizations as well as us. And then generally, you'll see a lot of examples of enterprise customers betting on us. From HSBC, ANZ Bank, that was there with me in the session. They talked about how they're getting value out of our data platform in general. It's amazing to see a lot more enterprises adopting and coming here telling their stories, sharing it with folks. Okay, Shadea, thanks so much for joining us on theCUBE. Appreciate it. Good to see you again. Congratulations. Data Fusion ingesting non-RAMPs into the superhighway of BigQuery, Big Engine, their large-scale data warehouse. I'm John Furrier, Stu Miniman. We'll stay with you for more coverage after this short break.