[MUSIC PLAYING] APARNA SINHA: Hi,
I'm Aparna Sinha. Welcome to day two
of Google Cloud Next. I'll be co-hosting today
alongside Urz Holzle, and we've got a really great
lineup of new technology and demos from Google Cloud. But before we get
started, I want to thank you all for taking
time out of your busy week to be here with us. Yesterday was an incredible
first day at Next. Thomas and Sundar made
some amazing announcements. This week we're releasing over
100 new products, services, and programs for you. We're kicking off day two
now with this keynote, a Cloud built for developers. After this, we'll jump right
into a live developer Q&A where you can ask us anything. And then tomorrow
is community day, and that'll be totally
dedicated to you for learning,
discussions, networking, and all kinds of fun. All right, let's get things
started by welcoming Urz. Hi, Urz. URS HOLZLE: Hi, Aparna. Hi, everyone. I hope you are as excited as
I am to get started today. And since this is a
developer keynote, we'll kick it off
with our first demo right away from Google
Research and our Deep Mind colleagues who've been
working on some amazing voice technology. SPEECH MODEL:
Welcome to Next 21. [SPEAKING FRENCH] URS HOLZLE: Now before
you get too impressed by my language skills, I did
not actually speak these words. What you just heard is a
custom text to speech model that has trained on
my voice and that can generate synthetic speech
in different languages. And our teams have
already started using this technology
to improve our voice experience for all users. APARNA SINHA: That's right. Google's project Euphonia is
already using custom voice to help people with
atypical speech to communicate and
be better understood. Technology like this creates
a more inclusive world. To find out more about
this, check out the project Euphonia website. Custom Voice is available today
for select Cloud customers. We're very mindful of
the potential misuses of this technology, and
we're taking great care to prevent them by reviewing
each use case uniquely. URS HOLZLE: Now look
around you, every day you see innovation that's
brought to you by developers like yourself with
persistence and skill, and Google has a long tradition
of supporting developers in open source and elsewhere. For years, you've been
using technologies like Kubernetes, Firebase,
Tensorflow, Go, Angular, GRPC, and many others. And when we built Google Cloud,
we built it for developers, and we were inspired by all the
things you've created with it. And our job is quite
simply to make it easier for you to do what you love. So we focus on making you as
productive as possible with the least amount of effort. And so in everything we design,
we take all the feedback that you're sharing
with us and build a platform that just works. Whether that's by natively
embedding key security or sustainability features
into the platform itself or by featuring
partner solutions right within our own console, we're
really focused on one goal, giving you the best developer
experience of any Cloud provider. APARNA SINHA: Google Cloud
has had tremendous traction with digital native customers
since our very early days. How have you seen
customer and partner adoption evolve since then? URS HOLZLE: Well, many
of our biggest customers are Cloud natives, but we've
seen tremendous adoption by a broad segment of
enterprise customers in traditional industries
as well like entertainment or financial services. For example, Major
League Baseball, which is North America's
oldest and most attendance professional sports league
is using Google Cloud to modernize fan
engagement and to increase operational efficiency. And Equifax, which
was founded in 1899 and is one of the
world's largest consumer credit reporting agencies,
is transforming itself from a credit bureau to a
next generation data analytics and technology company
built on Google Cloud. APARNA SINHA: We've
seen a huge shift. Essentially every company
is becoming a tech company to increase their
competitiveness and establish leadership
in their industries. Developer talent
and Cloud Services are at the heart of this shift. Whether you call that digital
transformation or something else, companies of
all sizes are finding that Google Cloud is
optimized to help make your data, your
applications, and your talent more useful and relevant
to your business. URS HOLZLE: Exactly. And as Thomas
mentioned yesterday, everyone needs to be
thinking through how they'll fundamentally shift into
a technology company to serve their customers in the
most meaningful ways 10 years from now, and you as developers
are key to making this happen. So organizations
ask themselves, do we have the most cutting
edge technology to become a leader in our industry. And you, the developers, are
well-equipped to answer this. And we are super
focused on making you and your company successful. And there's two
areas we focus on to support your growth,
first of course, making it easier for developers
to get their job done, second, investing in a
developer community so that everyone can learn
and grow from each other. So let's talk about
how Google Cloud is making it easier for developers
to get their job done. From our transformational
infrastructure stack to our deep innovations
in data, security, ML, every feature we release starts
with simplifying the developer experience. Take our open Cloud
experience, for example. We recently expanded
our compute stack to include [INAUDIBLE] VMs
that deliver 42% better price performance over any
other comparable solution in the market with no recompile. And thanks to our zero
trust approach to security, Google Cloud was ranked a leader
in IS platform native security by Forrester, well ahead
of the competition. And of course,
we're focusing also on managed services
that make it easy for you to deploy scale and
manage Kubernetes clusters on the edge or in the Cloud. And Google Cloud has the
most complete and most secure container experience
for developers. In fact, the 2021 Gartner
solution scorecard for Google Analytics
Engine gave GKE an overall score of 92, again,
well ahead of the competition. And of course with
GKE and Anthos, you can run these containers
anywhere on-premise, other clouds, on
the edge, anywhere. So it's fair to say that in
the years since Google invented Kubernetes, containers
have really completely revolutionized IT operations. Now recently European
filmmakers from Honeypot.io created a documentary on
the history of Kubernetes, and it will be out
in January 2022. And you are the
very first audience to have a look at the
trailer right now. So let's roll that. [VIDEO PLAYBACK] - Sounds good. - Do I look at you,
look at camera? - 2013, it was clear
that Cloud was a thing, but most folks were focused
on infrastructure Cloud. - The dirty secret for a long
time is like people who are either building their own
data centers using co-los, there's a huge resource waste. - And so at that
point, automation tools are all the rave. People are now trying to
abstract away the servers. - Google was looking
for ways to apply its internal infrastructure
expertise to the Cloud. - As we started looking at
technologies like Docker, we were like impressed
by the strength of what they'd
accomplished in solving a very specific problem. - This is going to happen
with us or without us. - Google had to make a bold
move in the Cloud space to be the long-term winner. - Every big startup, I felt,
had a container orchestration project, and half of them were
announced at Docker Con 2014. - Open source is most
successful when it's played as a positive sum game. [MUSIC PLAYING] [END PLAYBACK] APARNA SINHA: This is
such a great community. URS HOLZLE: Yeah, absolutely. I recognize a lot
of the faces, and I can't wait to see the film. Now let's get back to how we're
making it easier for developers for you to build really
the leading technology companies of tomorrow. Now Kubernetes
deployments can involve a fair bit of manual
configuration, clusters, nodes, load balancers,
animals, et cetera but not on Google
Cloud, because we offer you the most automated
and secure Kubernetes experience available. With GKE Autopilot,
Google provisions and manages the clusters entire
underlying infrastructure, including control plane,
node pools, working nodes. And that lets you
focus on the higher level of services
and applications that you're building. Nobody else offers
anything like this, because beyond
managing node upgrades, GKE Autopilot also automatically
configure security features like shielded GKE nodes, secure
boots, and workload identity. And it also implements
security best practices by blocking less safe features
like external IPS or legacy authorization. So you don't get
a toy Kubernetes cluster, which GKE Autopilot. You get a sophisticated
cluster that uses the best practices brought
to you by the team that brought to you Kubernetes itself. So you're always
up to date, and you get the same results
as the experts without having to be
an expert yourself. APARNA SINHA: The
pandemic put developers in the driver's seat, and
you all drove GKE usage to all time highs. At the same time, we
saw explosive growth in the use of Google Cloud
serverless offerings, especially Cloud Run
and Cloud Functions. It's mainly
enterprise developers who have driven this growth. Cloud Run excels at
developer experience. It's earned the highest
customer satisfaction rating among developers as measured
by user research international. Cloud Run combines the best
of both worlds bringing you serverless and containers. There's no cluster to
set up or configure, so developers are able to
scale seamlessly and securely. Under the hood, Cloud Run
scales container instances in isolated sandboxes. Any access outside a
sandbox is mediated by network controls or identity
and access management or both. And it isn't just for new apps. Cloud Run supports
traditional workloads like Java Spring
Boot and asp.net. We also recently introduced
committed use discounts to lower the cost at scale. And we've introduced
Always on CPU, which enables asynchronous
and background processes to be used on Cloud Run. So you have all the
benefits of serverless without the restrictions. The theme here is easier, more
secure development, especially with remote work. URS HOLZLE: You're
absolutely right. We've been focusing on remote
development for some time now, but the pandemic has certainly
accelerated the shift. Now what would be more
essential to remote development than to be able to use
the full power of TCP right from your laptop
with zero local setup? Cloud Shell Editor is a context
aware remote development environment that lets you
develop and manage applications securely from any browser. It supports languages like Go,
java, node, Python, C Sharp, and comes with an integrated
debugger source control API Explorer. And if you want to test
locally on your laptop, it also comes with local
emulators for Kubernetes and serverless APIs. APARNA SINHA:
Thanks, Urs. Next let me introduce you to Abby Carey. She is going to show us how
Google Cloud makes it easy for you to securely build
modern applications, again, right from your laptop. Hi, Abby. ABBY CAREY: Hi, Aparna. We developers have had a
hard time writing, extending, deploying, and
operating applications, but it doesn't have
to be difficult. Let's start with
Cloud Shell editor. It comes with current versions
of your favorite DevTools like Docker, Minikube,
Scaffold, and more. So it's nothing to download
or install locally. Tutorials are built
into Cloud Shell Ed, which makes it easy
to come up to speed on complex topics like GKE. APARNA SINHA: So no more
switching between tabs, docs, your terminal, and your code. This integrated
experience is highly differentiated
from other clouds. You can even offer
your own tutorials, and that allows
your organization to share best practices and
on board new hires faster. ABBY CAREY: Another
popular feature is Kubernetes YAML
authoring assistance. Let's say I want to add YAML
for a service to this project. I can press Control
Space and then find the Kubernetes service snippet. Now I can tap through
and fill everything in. I also get auto completes. And if I happen to make
a formatting mistake, I am notified that there's
an issue in real time. APARNA SINHA: Now
many of you prefer to work locally in an IDE. This same YAML
authoring assistance, it's also available for VS code
and IntelliJ via the Cloud Code plugin. Cloud Code has built in
support for both Cloud Run and Kubernetes. ABBY CAREY: In fact, if you're
using Cloud Run or functions, you don't need to know Docker. You can build and deploy your
app with just one command, because Cloud Code-- because Cloud Build is
integrated under the hood. This is an application
with no Docker file. With the new GCloud run deploy
command, all I have to do is provide a name for
my service and then let it know where my
source code lives, which is this current
directory, and we're deploying. APARNA SINHA: So nice. And thanks to this ease
of use, 98% of users deploy an application to
Cloud Run on their first try in less than five minutes. ABBY CAREY: I just showed source
code deploys the Cloud Run but there are more ways Google
Cloud has made deployment easier and more secure. First, I can scan my
build container images to check for vulnerabilities. I've already run an on demand
scan on one of my images using GCloud Artifacts
Docker images scan. Now I can copy the ID
of my scan and then view my images vulnerabilities
with the list vulnerabilities command. And once that's finished,
a severity level is assigned to
each vulnerability to help you prioritize. APARNA SINHA: That's
super important. It's really helpful in
addressing security concerns earlier in the software
development lifecycle. But now what if your build
pipeline is compromised? ABBY CAREY: For
that, I can enable Binary Authorization on my
deployed Cloud Run services. This way only trusted
container images are deployed to production. APARNA SINHA:
Binary Authorization is truly unique in the industry. It enables you to put proactive
security measures in place to reduce software
supply chain attack risk by blocking deployments
that violate policy. And speaking of
deploying, we're making it seamless for you
to do CI/CD securely. You can take advantage
of serverless build environments within your
own private network with Cloud Build private pools. ABBY CAREY: And
for advanced CD, we have Google Cloud
deploy, which allows you to create custom delivery
pipelines for your specific use case and needs. APARNA SINHA: That is so cool. Well, a real application
connects to many supporting Cloud Services. So Abby can you
show us an example of how we make these
integrations easier. ABBY CAREY: Sure. When creating a
Cloud Function, it's easy to integrate
with Secret Manager. First, create a secret
that stores your API key, which I've already done. Now I can either
mount it as volume or expose it as an
environment variable. I'll mounted it as a volume, and
then I'll name my Mount path. This will always point to the
latest version of my secret. And now I can securely
reference this API key from my source code. This abstraction enables
portability and a better local development experience. Cloud Run also integrates
with Secret Manager to make it easier to
do the right thing and not put sensitive
data in source. APARNA SINHA: Love that so much. OK, so now you've written your
app, you've deployed your app, and you've connected your app
to other Google Cloud resources. What's next? ABBY CAREY: Operating
your app in production. With Cloud ops, you
get one integrated view for your alerts, events,
metrics, and logs. No more jumping
around multiple tools as you try to understand
what went wrong. APARNA SINHA: That
was so awesome, Abby. Thank you for
sharing this with us. ABBY CAREY: Thanks, Aparna. APARNA SINHA: In each
of these instances, we've done the
integration work for you. Because the more work
we put into this, the less work you have to do. And this principle applies
to security as well. We've put a lot of energy into
building security natively into everything
we do so that you can innovate with assurance. Both GKE and Cloud Run
benefit from the security fixes we implement before
vulnerabilities are exposed. Just think about the
famous vulnerability uncovered in how Kubernetes
was handling proxy requests. We found it we coordinated and
communicated the disclosure, we fixed it for the entire
Kubernetes community, and we patched all our
products before any customers were impacted. More recently,
cyber threats have shifted the focus towards
the software supply chain. URS HOLZLE: That's right
malicious actors are trying to compromise the software
supply chain from bad code emission to bypassing this
CI/CD pipeline altogether. And to help solve
this problems, we proposed an industry
standard called Salsa. It's a security framework
that provides common criteria for increasing
levels of software security through automation and
through cryptographic signing at each stage of the
software supply chain. And that makes it possible
but not necessarily easy. And so making it easy for
developers to ensure security is super important,
and that's why we're focusing on
building the security right into the
developer tool chain anticipating and preventing
issues ahead of time, not when you're most at risk. So for example, Cloud
Build our service that lets you build,
test, and deploy across multiple environments,
such as VMs, Serverless, Kubernetes, or Firebase
now offers Salsa level one compliance by default.
Because Cloud Build gives you a verifiable build provenance. So this provenance lets you
trace a binary to the source code that it was built
from to prevent tampering and to prove that
the code that you think you're running actually
is the code you are running. Cloud Build is the first
and only CI/CD service to offer this capability,
but we go beyond that. As you've seen, build the
integrity automatically generates digital
signatures, which can then be validated
before deployment by Binary Authorization. That's another
Google Cloud first. And so without you
needing to do anything we prevent anyone
in your organization from deploying code that has not
been built by your legitimate build system. Now ensuring security
post-deployment is equally critical. On GCP, you can enable
continuous scanning, and you can use our service
mesh to embrace a zero trust security model and automatically
and declaratively secure your services and
the communication. So you can manage
authentication, authorization, and encryption between services
with little or no changes to the applications themselves. Let me say that again-- with little to no changes to
the applications themselves. So that means that these
security improvements help secure not just new code
but also existing binaries so you can use them
for any application that you're migrating
to the Cloud. Both Anthos Service Mesh
and now Cloud Build hybrid are available
across Google Cloud and your on-premise
environment, and they work with VPC Service
Controls and VPC appearing to automate the development
security for your enterprise. No other Cloud provider protects
your software supply chain to this level,
because we started working on software
supply chain security long before it was
in the headlines. And so by choosing GCP, you
benefit from this leading edge focus on security. APARNA SINHA: Whether
we're building foundational open source
technologies like Kubernetes or Istio or turning them into
fully managed services like GKE and Anthos Service
Mesh, our goal is always to reduce
complexity for our users by helping create these
industry standards we can provide safer and simpler
services for you, the developer and that's exactly our approach
to securing the software supply chain. We've co-founded the Open
Source Security Foundation with other technology leaders
to create security standards for open source software. And we're starting to
bring products to market like Open Source Insights,
which provides you a complete transitive dependency
graph for many Open Source packages. Now let's turn
back to Urs to hear why Google Cloud
is best positioned to support you in
becoming a technology leader in your industry
using data as a core asset. URS HOLZLE: Thanks. Yes, so far we've been talking
about developing and managing code, but data is at the
heart of many enterprises. So we also have the
leading data cloud products in the industry designed
for optimal performance and reliability for applications
of all sizes while scaling to immense capacity. Now let's start with databases. When it comes to databases,
every Cloud gives you choices. They offer SQL databases, which
are great but unfortunately don't scale, and of
course NoSQL databases, which do scale but,
unfortunately, are not SQL. Only Google Cloud gives you
a third choice with Spanner. Because Spanner is
SQL, and, in fact, it just got a
post-credits interface, but it scales horizontally,
and it can literally handle a billion requests per second. Nobody else has is
scalable SQL system, so it's no wonder we're
seeing huge adoption. Now on that data
warehouse side, we have, of course, the
leading Cloud data warehouse with BigQuery. Hundreds of customers are using
BigQuery at petabyte scale today, petabyte each. And you can run of course
BigQuery on AWS or Azure. On top of that
open source systems for data-like processing
like flank, spark, and beam run natively
on Google Cloud in a simpler and more
cost effective way than in other environments. In fact, you can
realize a 57% lower TCO compared to on-premise
data lakes for data science projects. Now on top of that
savings, our data cloud also includes the
world's first and only autoscaling and
serverless Spark service. And finally, Google
has deep partnerships with leading data driven
companies, including Data Flint, Confluent, MongoDB,
Reddy's labs, and many others. And so together
we help customers access an open platform that
powers analytics at scale yet is easy to use. APARNA SINHA: Our
partner community is central to the health
of our Cloud business, and we're especially excited
about the innovation coming from our data
Cloud partnerships. Together, we've optimized
our infrastructure for performance and efficiency
to give our partners that extra edge when
they run on Google Cloud. One of our leading data
partners is MongoDB. And we have their CEO Dev
Ittycheria here with us today. Welcome, Dev. DEV ITTYCHERIA: Hi, everyone. Happy to be here. APARNA SINHA: Dev,
one of the trends we're seeing in our
enterprise customer base is that they're now
competing for leadership positions in their industry by
becoming technology companies. How would you say Google Cloud
and MongoDB working together can help these customers
achieve that transition. DEV ITTYCHERIA: Well,
Aparna, the companies who are in the leadership
positions in their industries are those who have built their
competitive advantage using software and data to
transform their business. And the key word here
is built. You can't buy a competitive advantage. You have to build it. This means you need to enable
your developers to innovate as quickly as
possible, whether it's building new software to
seize new opportunities or to respond to new threats. MongoDB and Google Cloud
deeply understand this. Developers choose
MongoDB on Google Cloud, because we give
them the tools they need to be as
productive as possible, including having our services
available in the Google Cloud console for easy
discovery and deployment. Today MongoDB Atlas runs
in 24 Google regions across the world with deep
technical integrations with Google's
analytic and AI tools. This enables our customers to
innovate quickly and emerge as leaders in their industries. As a result, we're seeing
explosive growth or customers embracing the true value
of our partnership. APARNA SINHA: That's incredible. So when you think about
the development teams that these new customers,
what's the biggest challenge that you're helping them solve? DEV ITTYCHERIA: Yeah, when
you talk to development teams, you find that they
spend the most amount of their time trying
to work with data, as serving relevant data at the right
time to the right audience is critical to building
any application. Unfortunately,
relational databases are not designed for the way
developers think or code, nor are they designed for scale,
fault tolerance, or resilience. Consequently,
development teams find it hard to move fast using
relational databases. MongoDB is designed to
address this problem. We make it very easy for
developers to work with data, and we're able to address the
most demanding requirements for performance scale
and fall tolerance. The partnership with
MongoDB and Google Cloud enables developers
around the world to easily build modern software
applications to address their needs of
today and tomorrow. APARNA SINHA: That's terrific. Thank you so much for
being here with us today. DEV ITTYCHERIA: Thank
you for having me. URS HOLZLE: Yes. Thanks, Dev, for joining us. Now there's lots
of ways developers can improve their productivity,
automate tasks that are repetitive, master
the command line, use the best tools that
make your life easier, or reuse other people's
code just to name a few. And another great way to
accelerate your productivity is with building
blocks or templates or fully managed services in
areas like machine learning. Because on Google
Cloud, you don't have to be an expert to
build smart applications. With new services like Vertix
AI, you can build, deploy, and scale more effective
AI models quickly. So that lets you
deliver the insights to your organization that
will help them create more personalized
customer experiences, run more efficient processes,
and take that leadership position in your industry. APARNA SINHA: So with that,
let's go to our next live demo. Joining me today
is Anu Srivastava. She's going to show us how
these breakthroughs in AI are advancing Cloud adoption
and redefining the world of document processing. Hi, Anu. ANU SRIVASTAVA: Hey, Aparna. We all know how to
work with data when it's in a structured format like
in a database, some JSON, CSV files, or just variables
in my code, right. But what about
unstructured data? Many of the world's business
processes start, include, or end with a document,
but these documents can be difficult to process. Think about all the ways you
could enhance your application if you could just
unlock that data. This is where Google Cloud
document AI comes in. Doc AI is a platform that
has solutions and tooling for automating your workflows
backed by machine learning. We've bundled together some of
Google's flagship AI technology such as computer vision, OCR,
natural language understanding, and even Google's expertise
in building knowledge graphs all to provide you with
a simple yet powerful way to build applications
that better understand unstructured data. Let's go see a demo
of DocAI in action. So here we have a receipt. I was buying some
office supplies, since we are unfortunately
not back in the office yet. What I'm going to do is I'm
going to actually upload this into the DocAI platform. So in our Cloud console,
we have the DocAI platform where we have built
in preview mechanisms so you can test
out your documents. So this is going to
an endpoint, which has a specialized model we
have specifically trained on a variety of expenses. Google maintains and
improves the models for you. APARNA SINHA: Wait a minute. I hope this is not with my data. ANU SRIVASTAVA: Absolutely not. We never use your data
to train our models. Your data is only used
to serve your request. So let's take a look
at the data extracted. APARNA SINHA: I've
seen this before. Next you're going to
tell me that you're going to automate my expenses. ANU SRIVASTAVA: I knew
you would say that. This is the canonical
demo use case, right. But have you ever
seen it like this? Take a look at this field
that I'm highlighting, the supplier address. This address isn't present
anywhere in the document. APARNA SINHA: Wow. Where did that come from? ANU SRIVASTAVA: This is only
possible with Google's Document AI. The secret sauce here is
that the knowledge graph is able to not only give
you back the original text from document, but
it's going to enrich your response akin to what you'd
see in a search but as part of your API response. APARNA SINHA: Wow. That's really great. ANU SRIVASTAVA: And
it's not just this. We have several
specialized models for many more document types
of much higher complexity. Let's take a look
at this payslip. So I ran this earlier, and we're
looking at the preview output again. You can see that
we have some keys. We have some fields. You can see enrichment on the
employer name and the address. Once your data is in
a schematized format, meaning that we know for every
document of a certain doc type there are common important
pieces of information, so what we did is we
predefined a set of keys. So what we do is with
your extracted data, we merge your data to
these pre-defined keys. So it's much easier to
work with than raw OCR. So once it's in a
schematized format, it's easier to pass on
to a downstream service, or maybe you're using
something for analytics like BigQuery or Looker. APARNA SINHA: That makes sense. But what about ensuring
accuracy, and also do you have
multi-language support. ANU SRIVASTAVA: We know that
with important documents such as these, you can't
afford any missteps when it comes to accuracy. So that's why DocAI
provides a human in the loop configuration to trigger
on confidence scores, so either for specific keys or
on the entire document itself. And as for a
translation, we support over 100 languages, such as
Spanish, Japanese, Arabic. No other solution on the market
supports such a wide array of languages. APARNA SINHA: Human in the
loop, translation, and knowledge graph capabilities
that can be applied to a wide variety of documents,
this seems super useful. Of course, the next
big question is can it be applied to big,
bulky, complex documents like business contracts. ANU SRIVASTAVA:
Let's take a look. So here I read a contract
earlier this morning. You can see that there
are typical things you'd find in any contract. There are some document
names, the parties involved, some dates. And like with every
easy to read contract-- being sarcastic here-- there is an expiration term. So this expiration date
actually isn't present anywhere in the document,
and it's actually not easy to figure out. It's not in an easily
parsable format, shocker. Google's contract processor is
able to figure out this date value by understanding
signals found across the entire document. APARNA SINHA: Wow. Well, Anu, before
you go, can you tell our awesome
developers how they can get started with DocAI. ANU SRIVASTAVA: Absolutely. So I know we covered a lot
at just breakneck speed. So check out the
breakout sessions on DocAI to dive deeper. You can also check
out the documentation for code labs and quick starts. We have client libraries in
all of your favorite languages, such as Python, node.js. My personal favorite is
Java, but it's an API so you can really use this with
whatever platform or framework you're already using. We are thrilled and look forward
to see how you use Google Doc AI to power your applications. APARNA SINHA: That was amazing. Can I have a high five? Yeah. Thank you, Anu. I loved every part of it. ANU SRIVASTAVA: Thank
you for having me. URS HOLZLE: Another area that
Google has invested in deeply and that's becoming more
important to more companies is sustainability. Many Cloud providers have
a vision for a sustainable future, and many aim to match
their electricity consumption with 100% renewable
energy by 2025 or 2030. We accomplished 100%
renewable energy in 2017. So we're the only hyperscale
Cloud to do this today and all of that
with data centers that are twice as efficient
as the average data center. APARNA SINHA: This
past week Sundar talked about Google's goal to
enable over a billion users to live and work more
sustainably by next year. To reach goals like
this and those outlined in climate pledges made by
more organizations every day, we rely on developers like
you to do something about it, but we also know
that it's difficult. URS HOLZLE: That's right. One of the biggest challenges
that companies face is that they lack
the tools to account for environmental costs. And to help developers address
this for their organizations, we built sustainability tools
directly into Google Cloud. With Google Cloud
carbon footprint, you have access to the
energy related emissions data that you need for external
carbon disclosures in just one click. Now you won't need
this calculator if you just want to
report the net carbon footprint of your workload
on GCP, because on GCP it's always zero. We also have our
region picker where you can choose that data
center region with the lowest gross carbon cost right now. Of course, again,
your net impact is zero no matter
what region you pick, but this tool lets you
help go one step further to become carbon free,
not just carbon neutral. Now that's actually
a tool that I can't wait to deprecate
in 2030 or so, because Google
Cloud has committed to be 100% carbon free by
2030 every hour of every day. Now we also realized
there's still a lot to learn when it comes
to building sustainably. And to help we just
released a master class called Sustainable
IT Decoded with some of the world's top experts. So check it out for
guidance on how we can all build more sustainably. Now while we're proud to run the
cleanest Cloud in the industry, we're even more
inspired by the work that our customers are
doing with Google Cloud to solve climate
change challenges that are unique to their business. And today we bring you a
preview of Google Earth Engine and its integration
with Google Cloud. With over 700 data sets
and 50 petabytes of data today, Earth Engine gives
scientists and developers access to the world's largest
catalog of satellite imagery and the tools for driving
sustainable impact. APARNA SINHA: So let's look
at this in a bit more detail with an example of how Google
Earth Engine and Google Cloud enable customers
to assess risks arising from climate change. But instead of me
telling you about it, we've invited Joel
Conkling to show you. JOEL CONKLING: Thanks, Aparna. The world is
constantly changing, and that creates
opportunities and risks. Helping uncover critical
insights about the changing world is why we're integrating
Earth Engine into Google Cloud, and that integration is
now in private preview. Today I'll demo a
workflow that combines Vertex AI, Earth Engine. BigQuery, and
Google Maps platform to show how Google Cloud makes
it incredibly easy for you to innovate and deliver
insights and do it quickly. So here's a scenario. You work at an
insurance company, and you need to analyze
your company's exposure to flood risk. You think New buildings
may be a strong contributor to that risk, and you want
to test your hypothesis. To do that, we first need to
understand where the built environment is expanding. In other words, we just need
to categorize the surface of the entire planet. That could be hard, but
Vertex AI offers the tooling to develop a best
in class ML model, and Earth Engine provides
constantly updated data. Let's fast forward a bit. We finished training our
model, and now Earth Engine is sending satellite
imagery to be categorized, so your understanding
of the world can update in near real time. Here's the Earth
Engine script showing the results of that model. This area in red is
where the model estimates the locations of buildings. That's your current
built environment. To find the change over time, we
need a few more lines of code. These lines of code give us
a built environment in 2016. And here, we calculate
the difference between 2016 and today. When there's a change, it
shows up in purple on the map. This is where there
are new buildings. So next we're going to
export sample and export the data so we can do
additional analysis in BigQuery. Over in the BigQuery
console, this script clusters those data points
here and then outputs polygons that show the areas
with the biggest changes in the built environment. So at this point, you
have a few options. You could combine this
data with flood locations you identify around the
world also with Earth Engine. Maybe you want to enhance
your model with weather data and physical terrain data. That's available in
Earth Engine too. You could also include data
on your company's insurance portfolio to gain additional
insight into critical risks. We'll wrap up this
demo by visualizing our results and a new feature
available on Google Maps platform, the open source
data viz library deck.gl with a BigQuery connector
provided by [INAUDIBLE].. So we now have a clear picture
of where the built environment is changing and where to
focus next for our work on flood risks. In summary, no
wrangling data, no need to manage infrastructure,
just actionable insights incredibly quickly. We can't wait to see what
you'll do with Earth Engine's new integration
with Google Cloud, and with that, I'll
pass it back to Aparna. APARNA SINHA: Thank you, Joel. It's incredible to see
how our customers can use our sustainable technologies
to address climate change now. I'm really inspired
by all the things we've talked about today. And thinking about
how you're going to lead your companies into the
future, that's super exciting. No pressure, but it's
really up to you. We've invested millions
in the developer community over the last five
years and we'll continue to invest in the coming years. And Urs, as proof of
that, I understand you have some additional
news to share today. URS HOLZLE: Absolutely. I'm really excited
to announce today our new developer community
program called Google Cloud innovators. I want to welcome and introduce
our first group of leaders who are driving meaningful
impact in the industry and their communities. So take a look. [VIDEO PLAYBACK] [KEYBOARD CLICKING] [MUSIC PLAYING] [END PLAYBACK] APARNA SINHA: This
is so exciting. URS HOLZLE: Yeah,
through this program. We'll give developers access
to early technology previews and Google engineers. We'll recognize the expertise
of our community influencers by promoting their
contributions, and we will work
closely with them to solve the toughest problems. So we're excited
to come together with this group of innovators. Join us at
cloud.google.com/innovators. APARNA SINHA: So cool. I've been waiting
for all this time. Community is extremely
important for companies to create that much needed human
connection with developers, and we hope that this
gives you a window into the motivation
that you all give Google to build Cloud products
and services that developers love. URS HOLZLE: And we look
forward to partnering with you to become the greatest tech
companies in your industries. APARNA SINHA:
Remember to join us next at the live
developer Q&A session and also tomorrow
at community day. Enjoy the rest of the show. URS HOLZLE: Thanks, everyone. [MUSIC PLAYING]
