Accomplished product management leader with over 20 years of experience building industry-leading products at startup and blue-chip companies. Proven record of leading product teams to deliver business results through customer insights and innovation.View the profile
Rare combination of technical, business, customer-centric, and evangelism expertise. Passionate about solving customers' exa-scale analytics problems using innovative and emerging technologies. Experienced with growing business from X M to X B.View the profile
About the talk
Today’s enterprise demands from data go far beyond the capabilities of traditional data warehousing, and for many leaders, the need to digitally transform their businesses is a key driver for data analytics spending. Businesses want to make real-time decisions from fresh information as well as make future predictions from their data in order to remain competitive.
Learn what’s new in BigQuery, Google Cloud’s enterprise data warehouse, and hear about all the latest feature innovations and see some amazing demos.
Speakers: Sudhir Hasbe, Tino Tereshko
Google Cloud Next ’20: OnAir → https://goo.gle/next2020
Subscribe to the GCP Channel → https://goo.gle/GCP
product: BigQuery, Looker, Sheets; fullname: Sudhir Hasbe, Tino Tereshko;
event: Google Cloud Next 2020; re_ty: Publish;
Welcome everyone to what's new in Big Bear recession and today has a and the director of product for smart and Patrick platform at the Quarry. Abstract flower in the smart analytics. Team vision is to give you a provided you with an enterprise-class data platform and I believe in telling his last Field Services meet all your security and trust me. So basically I have been focused on that in addition to that we are also about being open and flexible choice. And finally and most
important we want to embed intelligence across the whole platform so that it's now it's 10 years old in, June be celebrated at 10, that was pretty has customers from a few terabytes to hundreds of paradise, is it completely surveillance technology? What do I mean by that means you don't have to worry about any interested. Stop analyzing the new monitoring. There is no Administration, there is no machine. There is no note that I know servers that. Most of that is managed by by
Google. I mean just tens of millions of events in real-time into Bitcoin and north of organizations. Like I like kings of the gaming company, they have in game events that happened at the time so they can analyze what what users are doing Monday. So it's possible. Can you ingest large-scale events? It does not impact your credit performance, on top of a badina, because the complete distributor compute and storage. In addition to that, we have embedded machine When you can find SQL commands to create machine learning model.
Predicts inside the organization, walk-in signs for every one of you made a lot of investment into space. When do query is being used by organizations like a ratio, but they were able to do that to everybody in front lines that they can make better decisions. Play baddest kerbler. Powering some of the largest of the double close in analytics-based. Largest customer has more than 350 Paradise Abita in Bonita Warehouse was roughly hundred trillion. Bruce of Deborah Credit in a single query
and 10000 concurrent users being used by customers bank is highly regulated, keeping Stephen able to do. I walking with HSBC is a white lot of security capabilities in all of these to ppsc all these capabilities with HSBC. They're proven there are there are used by some of the largest banks in the world energy and bring dinner from 40, + countries or transaction data data data warehouse, they analyzed all of the data. And at the end of the day, Regulators about the viability of the business and all of this is running a few
minutes. So much more than a data warehouse because we have so much of these capabilities, smart analytics, platform and multi services to gather with the Quarry and sky is one of them communicate directly to Pub sub data for real-time analysis. Able to visualize with customers are doing what they're watching, their preferences are able to tailor customized. Contact advertisement. And if you look at the growth of bickering, like we've seen tremendous growth in last last year or so, are we have customers across every industry?
Vertical across every region in the world near blast, customers like Home Depot to UPS in in are in United States. Where is HSBC near Skyview King? We have a Metro Car Auto in in m, e r b, e r customer, send a Jeep, a creature like Fast retailing group are in the bank and go check and aeration in adoption across the globe. I buy different customers. The different kinds of use cases and highlight in the roadmap and able to share them with you and intelligent. And and
Let's start with prudence and every so we really need to think about it is when chance dependability and four areas availability and administration and control performance and Aquarius and finally trust security and governance. Today. Announcing the fact that we are enhancing rsla by an additional nine. So now there's four nines availability of our customers. There's no special race. They have today against a specific circumstances, it is actually industry-leading when it comes to Cloud. Our warehouses.
And of course, we care about performance and our customers care about four years. To demonstrate the power cord is a quarry on entire petabytes. Worth of gas is Cory later this quarter in Just 2 minutes and the best part of all is that you didn't need to do anything special to pay anything extra to start getting stuff for you. In fact, just in 2020 alone from the beginning of the year, with made major improvements to all the typical industry-standard Benchmark. This is a chibi
CVS and we're looking at the gyro mean performance. Servo for 1 and 32 scales from twenty to thirty percent. And again, you don't have to do anything and it just happens. Show me the number of customers in all content areas and of course, and sports teams are taking advantage of the journey that MLB took all the key stakeholders would be better than the previous form. And not a lot of things with them to worry about. So we talked about how do you sterilize it doesn't require Administration. That's the more customers we have that multinationals are complicated
asking for better control. Reservations is an enterprise-grade work with no power that allows you to Superstition because it's using Auto capacity. Enhancement is flex, lost allowed to purchase capacity. You since we get in some capacity to meet your work, Classic database administrators out there. I would need a number of enhancements as well as information. We can get him happy. Delighted about the seizure is coming in the next few months. Free slot auto scan, disk
slot, space pricing and you get like the management overhead model Kailyn. Let me see your spelling and the auto parts in. So we can go from zero to full-scale. And back to see if you'd like that because we are not power by the we don't have a supreme warm. When you're not getting optimal perform the second you get your capacity for fall into that 100%. Also, if you have Cory's in flight or is there being You can add capacity and no squeeze would immediately start taking advantage of that.
Right? So you can accelerate course in flight. Because we don't have gas. We're not powered by. Look we're not power by local desk, right? So we don't you start running our local, did you start getting performance Clips around the document? We don't have such problems right? We always have hot data for you. And finally I'm kind of already talked about the semester but we can provide a high level of reliability because we're immune to downtown. The next teacher were launching in the ministration. Next 2 months is admin panel. What
Native monitoring capabilities will provide you with, this will have a real-time components which you can see what's happening right this second and can go back in time, several weeks bus. So how are you going to use that? You going to be able to figure out how your organization using recording Channel. Customers always ask her, you have oldies, but how do I choose? Which model to use? Which one is the right? One for me. And you know, how how much capacity to
send you inside automatically actually recommend? What do you see on the on the damn bottle? Or was she was whether you should be on the spot, and of course we're going to do this without sacrificing. Going to tell you that it's going to recommend you to save money without extra sacrificing performance. Performance important. Aspect of Corey, is it's during our customers trust us with their data, and there were closed, and we can't stay on them. So, we have a kind of a
Wrong strategy here. The first one is Google employed, thousands of be best security engineers in the world. They are the working on security Hardware, networking penetration. So on so forth and before he gets the benefit from all of this innovation. That's the first killer. So you just get a high level of security automatically recorded the second giving you a number of controls execration and services that I'm able to customize how much security want in your vile things. I can send you a message and customer is a catalog.
Catalog is a metadata management and Discovery service that we launched recently cloud and the Donner Pass through data loss, prevention a t, i g. I Nation the private sphere formation and then you can apply tags and policies and Grant access to the right levels of data to the right, two users inside of your organization has a super flexible, super powerful platform that allows you to Dempster pillar as critical to Our Success here is the fact of a working with all the relevant agencies to achieve these artists A certification.
Right? So, you know, you can trust us because we have to give a certification with you store in financial data inside because you know, you can trust us because we fed ramp, we can do business. That's now. Now we talked about to and dependability, let's switch gears, talk about the intelligent key capabilities, and who will flower is a course, out, our capabilities are on machine learning and AI. I need to talk more about how we are enhancing at embedded machine, learning
capabilities that you're fighting with the credit. For. Let's talk about democratizing inside Devon, you fill us in Google Cloud. He is the one in size for every one we want to make sure every individual person in an organisation in this realm. The first thing that you're doing is on our research too, because you are a lot of our customers have been asking for capabilities in the, I'll be riding. 44 Dyckman can also write a love Lucy, equal and create streaming Point lines with pops up and data
flow and combination with Quinn. So I want to complete or suggest I water the table that I should be using and stuff like that, and investment going in that space. And the second thing I'm super excited about this feature, which is connected, sheets, sheets and literally Singapore, Google sheet in front of liquid. You can have one of my skill sets and Equity, you can use sheet to go ahead and create pivottables, jobs are functions in a custom calculations on top of the sheet and then all of that actually runs on top of the credit. So it's the best of both worlds. They've
experienced with blue. I personally use this everyday, I have a 29 + billion room table, so severely make data at scale, accessible to all analysts. Who can you stretch? It has a really powerful country. If you don't want to stop at Anna, listen and then and just a few books and organization Exchange in size to everyone else. In the organization with that, your beauty is natural language. Increase or decrease you literally. Go ahead and ask a question, get an answer.
It's asking for the back, there's no setup required on top of a p. I pick the table that you want to go to ask questions on and we just make it possible for you so you can just start We also believe that no one interface is the right interface for a for technology like this right now. I think organizations want to build custom work experiences on top of technology. So we do have the whole API organizations. Can people be integrating it into Google Sheets with connected sheets? Plus, DirecTV now have a very powerful
experience providing experience will be working with Denise other experiences that we can provide. All of this is actually the clear paper that you can read more about Google Assistant. Use analyze that gives us the ability to go into question. Palestine intent, identification matching because of Google's investment in the space of natural-language conversational, understanding and leveraging without preparation and said one of the key values that we haven't the credit email. Want to bring computation to data now. If we don't we don't like
coughing there that start the video of Running Scared. IMDb When did Kerry, Washington, do with two lines of sequel to create machine learning models of time, whatever you want? And then complain. When we launched couple of years back, we only had regression and logistical and linear. Regression model, support payments plus train station in a scene or how can I use the same combination of the new customer segmentation using Caymans Place train station to find products? That we should recommend a lie, basically. Email campaigns and our camping, so, very common pattern.
In addition, to leave also launched a Time series forecasting. This is great for inventing optimizations and stuff like that. So interested in this morning, Most important to us as creating them easily and running back to work. But I would love to use these models in real-time pocket Edition, how to re-enable that and to be able to export the model. And so you can export the model and then hosted in any of the serving platform that supports which would be Are we that we recovered the democratizing
inside at 7 a.m. connected sheets or embedded machine learning the total population that we have spent significant amount of time. As I mentioned, your customers are putting in like tens of billions of events for second event. And just this year, our whole storage engine from like a bath. First motor, screaming motor screaming first storage system now and you didn't seem to do any of your event or second without any any problem to more important matters.
None of us will get any have any impact on performance because of that injection break. So we can give you the same quality performance in the real time. Do you know when you have none of these data changes happening batch out of steam, you want materialized? So that you can easily power the dashboards and reports that you may have. All of this is all the way in the materialized view audio. What that means says, you don't have to. If you don't want to update me to the basic cable bill, always keep it on.
And most important is very real and Vivid automatically, think that query query passing engine will be lines. That you have a new materialized view which can make the career and faster and it automatically leverage that I'm not there for you to use. Now I'll be in August 1st, August 1st, a b, i n g o b. I n g. Working between us to deal with the technology. Make sure if we're going to keep him happy, but more importantly in this year we're going to extend that paper with your bi engine in memory cache
on the Quinns in between even from a standard aquarium in addition to that. And also all of our be our partners, like Tableau liquor license, do most of our be at all of them. Be able to leverage, the wi engine size of fascism, everything. And then Ending. If you really think from a stock perspective, us customers too late. Now, have a choice, you can decide whether you want no leniency, or you're okay with the Hyatt Regency and you can decide how much you want to use on the very hard data with cashing with the engine, on
the time tables or desk or the storage choice that you have, but you ride the query leaving, automatically first try and be an engine, not there, and plus spring. If not go to the core storage and the best price you can choose how to watch mechanical. Intelligence and Andrew and dependability, let's switch gears and talk about open and flexible design of the big thing for us. I will cover like, Big Bang Theory. We're flexible pricing as well as what is the cranium mean?
We have this vision and in very, very like to bring computer today. And not have customers have data not in GDP per capita in AWS S3 and other other customers have an answer. So with the only available for AWS, we will allow you to write queries and executed on top of the data is stored in Sp. What is this mean? People literally run out of fuel injector. Competition engine obituary in AWS data. You don't have to worry about paying for the secret charges or anything you can stop writing for the
same people that we support as well as you know and run it on a w. S o n be able to do competition in Dothan and then get access to the two or three different kinds of use cases that are becoming really as bad as this is going to allow you to have a single interface for a drone in computer so that I can extend it to that. Part II news came that it's becoming pretty prevalent have been talking to a lot of customers who have Hundreds of coronavirus in an S3 by
answering this case, what we can do is we can enable organizations to do data preparation summarizing the identifying the key there and then just moved on. And then when does machine learning model to advance and see how we have been learning as we go, and would love for you to request access, get access to this technology and let's talk about pricing. We talked about how Corey has some of the largest youth cases in the world. But what's interesting is before you probably also has some of the
smallest use cases on the world. And it's really because our system to be scaled area until the right pricing options for. Kind of scale. We're going to have the right price, Missouri. Start caring about additional capacity. Price predictability and you're working on these type of annual budget. We have flour, a pricing for you, to wherever you are. And, of course, you're not just using depilatory, typically, our customers are running spark, or closed out of pandas dataframe,
other types of execution engines. In other places are open. So we provide allow you to run sparked by producing other open source. Workload directly on top of the Quarry story at I levels of Troopers and parallelism so it's running. Open storage but your data lives outside of the Quarry, we provide that capability of federation allows you to in real time access data and relational databases from McCrory execution on using your phone and of course we support.
Finally, your daughter may be living in these other service providers, like stripe Marketo, maybe have Google internal data sets and live with an ad words with an YouTube and show some support. How do you get that data in our house? So you can take care of the rest once your connection is set up, we simply moved out into the quarry, on your behalf and now you can do is upset or summary of all the different, different capabilities that we are announcing.
There's not enough time in 30 minutes to 35 minutes for us to cover, everything stations that are happening, so that you should be able to get into all of these different topic as well as hear more about aspects of the How do you say be quiet? Is a providing the lowest total cost of ownership for our customers across the globe have been looking at the study and then trying to carry out and then finally we are the leaders in in on the different magic quadrant from doctor as well as far as the
last couple of years, is one example of a play, a body telling you is when covid-19 head, I know if everybody is is, is is fighting came to us and we were able to go out and make 20 + data sets available. If you are able to go ahead and make the competition on top of that, I think I'll leave the key thing is to be carried us for wine. Has this platform that we can share the love, share insights with the globe, not just with an organized. An industrial.
Buy this talk
Buy this video
With ConferenceCast.tv, you get access to our library of the world's best conference talks.