[MUSIC PLAYING] FRASER MACDONALD: Welcome. Thanks everyone for
coming this afternoon. My name is Fraser Macdonald. I am a product guy on
IoT Cloud at Google. JOHN HEARD: And I'm John Heard. I'm CTO at Smart Parking, a
company that's using IoT Core. FRASER MACDONALD: So thank
you all for joining us. We want to ask you
a question just to get things rolling
here, just to get a bit of a sense of who's
joining us this afternoon for next 45, 50 minutes. So just by some hands, how
many in the crowd today are working directly
for a municipality? OK. So we've got some hands
for people that are working with cities themselves. How about people that
are working for vendors that serve municipalities? OK. So a good number there as well. How about people who
have lived in a city? [LAUGHTER] Oh, what do you know? All right. So we've got some subject
matter experts here. That's great. And I say that kind
of half kiddingly, but one of the points that John
and I want to talk about today is a bit of an idea
of who are we building these smart cities for? What is the reason
for making them, not just how you make them? So I think everyone
in the audience has got a bit of
a voice in that. So thanks for joining us. So let's warm up with
a bit of a teaser. Take a look at this kind of
echocardiogram-looking thing off to the side here,
and see if you might recognize some patterns in it. Does anyone have any ideas of
what we might be looking at? Yup. Yup. Some urban areas, some parking. They're definitely urban areas. You're right. In fact, if we show
the names here, we're looking at
some familiar areas. You got the Bay Area off to the
side, New York, London, Tokyo, San Francisco. So what we're looking at
here is movements of people on their mobile devices. Every one of these
dots is somebody that has opted in in Google
Maps on their phone for location tracking. So you're taking that data,
anonymised it, collected it, and you're looking at the
living, breathing city here. This is a 24-hour
cycle of where people are going, what
they're up to, and this is enormously valuable data. It gives us a bit of
a sense of context for what we're
talking about today. There's an incredible
amount of information that's available to us,
if we can make use of it. So this is for people's personal
uses on their phones, and maps, and seeing where they've been. But imagine how
useful data like this could be to people who are
planning traffic patterns, people who are trying to
plan pickup routes for waste management, somebody
who's trying to get a new freeway
going, somebody who's trying to choose where the next
neighborhood the businesses should grow in. We're seeing a growth
in mobile devices. We're seeing a growth
in IoT devices. And we're seeing a
simultaneous growth in the ability to process
this data on a massive scale. So this is just the
beginning of smart cities. We're getting into a
really exciting age for it. So a question-- you
might have expected coming to a tech conference,
coming to a Google conference, that we're going to talk
purely about technology. And is this what "smart"
cities mean, the technologies? So, yes and no. Yes, in that these
are definitely ingredients of a smart city. I mean, having access to Wi-Fi
is what gives us the ability to collect data. Having RFD transit passes
are fantastic resources for municipal planners to
know when and where are people going, so we can
make smarter routes. Having self-driving
cars are going to be an enormous
overhaul of what it means to live and
be transported inside of our cities. But these are ingredients. I mean, in the same way
that an ingredients alone don't make a cake, this
isn't the smart city itself. What John I want to talk about
today is a bit of a shift. And we're wondering if
you'd join us in this shift from thinking about
a couple of nouns, a couple of the technologies
of smart cities, to a few adjectives that
might describe them. So, over the next little bit,
we're going to do a few things. One is I'm going to share a few
stories of what we've learned at Google from some
projects that we've been involved in that overlap
with the smart city space. So a bit of stories that
we've had some learning from, and a bit of what-ifs. And then we're going
to hear from John. And John's going to tell us,
well, if that's the what-if, this is what it
actually looks like. John and his
company are building the kind of infrastructures
for cities around the world right now that let them
make good use of their data. And we're going to
wrap up with a Q&A. So before we dive into
some of these stories that we've learned
from, I should call out that this session is one
in which we have Dory. So you might have
been in a session this morning that has as well. It's a Q&A tool that
we've got at Google, and we're rolling it out and
trying it at this conference. So if you go into the
mobile app that you have for this conference,
and you go to our session, you might notice that
there's a Q&A portion in it, where throughout our talk,
if you see something that looks interesting, or a
question crosses your mind in the moment, you can
write a Q&A question, and the rest of your
fellow audience members can then vote it up. And at the end of
the session, we'll mix a bit of microphone
questions and Dory questions from our moderator who
will read them out. So if a question
crosses your mind, feel free to hop on to your
app and ask it as we go. OK. So let's take a look at
some stories of what we've learned at Google from
some smart city adjacent and overlapping projects. What's the first adjective
that comes to mind? The first adjective of
smart cities that at least John and I when were sitting
down to chat about this thought it was efficient. And I think there's
a reason for that is because it comes up the most
regularly for smart cities. Take a look at this. This is a list of the most
common smart city projects that are under way
right now in the world. And if you're
taking a look at it, you'll probably notice a bit
of a pattern to it as well. You've got parking, street
lighting, waste and energy management, and then a whole
bunch of transportation logistics and optimization. These are efficiency problems. And one part of
it might be yeah, these are the low-hanging fruit. These are areas where an IoT
sensor or actuator can really help and not be too
complex of a rollout. But another big reason for this
is these are areas of high ROI. Each one of these
is making the city better able to spend
its money or better able to allocate its resources. And a really important
adjective of a smart city is one that is not wasteful
with its resources. So what does an efficient smart
city look like in the project that we've worked on? Some of you might remember
this demo from Next 2017. This is a bit of a what-if. This is something
that we showed, which is taxi cabs
driving around the city. So we've got some
simulated fleets of taxis across New York. And these taxis are
kitted out with something that's a little different. And instead of having
just a display ad in the back that's showing you
a clip or an ad for something, they're also location aware. They know where they're picking
up, where they're headed, what time of day it is. And with just a little
bit of that information, we were able to personalize
the ads to whoever might be riding in the car
with an incredible amount of accuracy. It's a tiny bit of
additional data, but it produces a huge outcome. So this is an example for
ads, but imagine this kind of intelligence if it were
applied to real-time traffic control. If your city, or you're
building a tool for a city, was able to equip the city's
traffic-control measures to respond to floats in
parades, to respond to an event that spontaneously happens,
to respond to a live traffic conditions. Or imagine if this
kind of real-time data, and a little bit of
location and time awareness, was able to affect
waste collection. So that's exactly
this next example. This is a company called Enevo,
and they're a Google customer. And they're building
something pretty cool, which is a waste management solution. And instead of being
the traditional take on waste management, which
is usually sort of schedule oriented-- you know, you know
your time when you're going to get a
pickup, truck comes by, they pick up your bin
whether it's full or not, and they drive off-- Enevo decided that
they were going to add some sensors and a bit of
an intelligence to the process. So they're parking-- or rather,
their waste management solution is aware of how much
trash is in the bin. They sell this to companies that
need a more efficient pickup schedule. And all they did was
put a couple of sensors in to know when trash
is getting full, and slapped a GPS
on their trucks, and out of a small
change like that, their customers
got better service because they had reliable
pick up and didn't have to think
about scheduling so much as starting to think about
waste management as a service. And they themselves got better
optimization out of the route because they could tell their
drivers exactly where to go and where they were needed. So what's another adjective that
might give us a bit of a place to start for smart cities? Participatory. So what would that mean? There are a couple
of perspectives you could take on this. And one of them is
participation of the people that live in the cities themselves. So here's another project
that we had some overlap with at Google. Some of you in
the audience might recognize where in the world
we are flying over right now. This is Japan. And specifically,
this is an image that was captured
back in 2011 when in the Fukushima
area of Japan there were three nuclear
reactors that melted down. And it was a major
public safety concern. And there was a pretty
incredible response in 2011. What we're looking at here is
a couple of people in Japan that were citizens put
together Geiger counter kits. They were makers and
tinkerers and shipped them out to people, other citizens
around the country, who got them, built them,
connected them up to the web on a platform called
Patch Bay at the time, and overlaid it on
top of Google Maps. And what you're
seeing here is people getting involved in
producing and engaging with the data of their own
cities to make them smarter. This was real-time intelligence
on, should I be in my area? Am I safe? Now, this is an interesting
and personal one for me because this
platform, Patch Bay, went on to become the Xively IoT
enterprise platform that Google just bought earlier
this year and is getting rolled in as part of
the IoT Core Cloud offering. The next one here
is Sidewalk labs. You might have heard of this one
as a sister company of Google underneath Alphabet. And Sidewalk has
a mandate to not think about the city of next
year or three years from now. They're thinking about
what might cities look like 20 or 40 years from now. So this is another project
we're involved with. It gives a bit of an idea
of where might we be headed. And Sidewalk's taking on a
really ambitious project. It's another one that's a
little bit personal to me, as I was in Toronto
by myself, growing up not too far from that
watercolor toothpick on the left there. This is a micro-city
that Sidewalk's taking on in Toronto. They're going to wire up
absolutely everything, from the municipal
infrastructure, to the transportation, to
waste management, to lighting. And they're going
to try to figure out what can we do with the data. So this story is just beginning. The Sidewalk story
is just starting. But what would it
look like if we wanted to try to effect
some of these changes and play with some of
these ideas right now? We've done some of these
projects as Google. What are some of the tools
as GCP that we're rolling out so that we can help companies
and municipalities try some of these ideas themselves? So how might I tackle efficiency
in a participatory city? Let's take a look at Google
Cloud IoT for a minute. So like much of
GCP, we're building on the exact same
tools and stacks that we build on for
our own products. When you're looking at the
Google Maps infrastructure, you're looking at
the same company that's producing GCP tools
for your company to use. And we took the same
approach with IoT. This is an example of what
an infrastructure might look like for powering
a smart city project, using a couple of
GCP building blocks. So on the left, we've
got our devices, whatever they might be, whether
those are occupancy sensors, they're smart
energy meters, they are watching out for
traffic patterns. These are the
devices, on the left, that have something
to say or that might be controlled like a stoplight. And the tool that we put
together in the middle here is Cloud IoT Core. So this is a recent product
we've started rolling out specifically for this use case. There are a lot of GCP ingestion
tools for huge amounts of data, like Pub/Sub. But the unique thing
about Cloud IoT Core is that it's also
meant for talking back, and it's specifically
meant for devices. So this is something that
can run with any hardware that your city is
bringing along, any sensor or actuator, over a number
of different protocols. MQTT is a lightweight protocol
for devices that are a bit more resource constrained,
HTTP for something that has a bit more compute
power and overhead and doesn't need quite that. But this is a way for
you to give your devices a unique identity, and be
able to talk back and forth with them, and tie them into
the rest of the GCP ecosystem so you can start making
use of their data. And we're going
to talk a bit more about the rest of
this picture of what is making use of their data
look like just in a minute. But we're looking at here
is a bit of add these blocks together and we can start
playing with the ideas that we were looking at in
a couple slides earlier. All right, let's take a look
at a couple more adjectives. Responsive. So responsive
seems self-evident. You would want a
smart city to be one that can correct if
something's going wrong, something that's aware of what's
happening in its own grounds, and something that
feels like it's not just operating on autopilot. What's an example of
something responsive? Well, here's an
example of something that we've learned a
bit from at Google. And this is a parking feature
that we've rolled out, which is based entirely
on machine learning. This isn't the same
level of granularity as what John is about to
talk to us about in a minute. This is just high level. This is just a
bit of information on where people are
sometimes and trying to make some estimations on
traffic patterns and parking patterns based on that. So if I'm planning
a trip to the mall, or I'm going to
city I've never been to, I can get a bit
of an idea of, hey, it looks like I should go at
10:00 AM instead of 4:00 PM. And this is an
example of something that doesn't just show you
where the parking lots are. It's responsive to when
you might want to be there. What's another adjective? Self-aware. So this is one that I find
really interesting and really important in the
smart city space. And its core to rolling out any
new programs as a municipality, or its core to being a vendor
to the cities, because you can't make good decisions
unless you know what's going on your own grounds. What does self-aware
look like for a city? Well, a lot of you I bet
have used Street View before. I was just doing a
trip last weekend and had never been
south on Highway 1 before and was using Street View
to check out a couple of towns and where's a nice
spot for a hotel. So, that's an example of how I
can get a bit of an awareness as a consumer. But one thing that we learned
at Google from Street View was the incredible
richness of image data. And this picture that we're
looking at was useful to me as somebody taking a
ride, but how could it be useful to a city? You might also be
familiar with Google Lens as a consumer,
something else that makes a bit of
intelligence from images. This is an example of, I
take a picture of a flower. I can have that flower
identified for me. It was something that
rolled out at a recent I/O and is chasing that exact
same problem of, well, we've got a lot of image data
in cities these days. Is there anything
we can do with that? So this is a consumer example. What might that look like
in a smart city context? How can I take these adjectives,
responsive and self-aware, and look at some
Google tools that would let me chase that, that
would let me learn from that? So I'm pretty excited to be able
to chat about this one today, as it's a relatively new
idea that some of our teams have been playing around with. And it's a combination
between some of our learnings from
Street View, and our machine learning and image
processing tools, and the GCP tools that
are offered publicly. So none of this is you have
to ask for special permission. This is all GCP
tool capabilities. But it's an idea that
we're looking actively for early partners on. So if it looks like something
that's interesting to you, reach out to me after the talk. You know Lens is the
consumer product. It will take an image
and tell you a rough idea of what's in the image. What you might also know
is that inside GCP there are a couple image processing
tools that go deeper than that. The Vision API is effectively
that for your apps. If you're building something
that has a photo capability, you can send the image
to the Cloud Vision API, and it will return with
a level of confidence what it thinks is in that image. And that's great for
a lot of use cases. If you want to be able to
tell, yeah, this is a car, or this is a park,
or this is a tree, it's an incredibly
powerful thing to build in with absolute
simplicity to an app. But what if I wanted
to go one level deeper? What if a city
really wanted to get to know what's going
on on its grounds without having to kit out
an enormous sensor network? Well, let's go one level deeper
to the next product called AutoML. And this is where you can
start training your own image recognition models. So this starts to
get really powerful as cities start wanting to
ask questions about what's going on our grounds. Now, our own Street View
imagery isn't publicly available and isn't available to process
for a number of privacy reasons. But images that a
city is taking itself are well within their
right to process. So, what might it look
like to take a look for, are there trees close
to power lines that might be a risk right now? Are there potholes in
the road that we might need to be keeping track of? These two products are
built for exactly that. Cloud Vision API
gives you a good idea of what's in a picture and
is dead simple to integrate your applications. You pass it an image. It passes back what it
thinks is in that image. But if you want to
go one level deeper, and you're working on an
application for a city that says, we'd love to help you
answer a very specific problem, then Cloud AutoML will let you
very simply, without a data scientist on your team,
create new models that will allow you to identify
what you're looking for. So imagine what this
can do for cities. If you could get a better
idea of real estate value out of images. If you could take
a look at where are we at risk for a
tree falling on a line? Where are we looking promising
for new areas of growth based on businesses
and buildings that are booming in that area? An idea of, could we
look for indicators of energy surge and production? Could we look for measures
quantitatively of development, of poverty, of living
conditions in our cities? These are tools that
are on GCP today. So, we've talked a bit about
a couple of the projects that we're working on at Google. And we've talked a bit about
what we learn from them, and what are some
of the GCP tools that might help you ask those
same questions in your cities. Let's go one level deeper. Imagine if you were never
stuck looking for parking. And to answer that
question, please join me in welcoming up John. [APPLAUSE] JOHN HEARD: Thanks, Fraser. Yeah, we're living in a
changing world, aren't we? And parking is one
of the things that we live with which really sucks. And so, I just want to tell you
a little bit about who we are and what we've been
doing for now 15 years. We've had to build a lot
of technology ourselves because there are no standards. But we're doing it in the
real world, everyday, today. So, we're based out of several
locations around the world-- Australia, New
Zealand, and the UK. We actually dogfood our
own technology in the UK, and we run hundreds and hundreds
of carparks across the UK with our own technology. But in New Zealand and Australia
and other parts of the world, we have a number of very
significant smart parking systems in place, tens
of thousands of sensors. We build sensors that go in
the ground just like that. That's an example of one. It's not much bigger
than a D-size battery. And it has to last for up
to 10 years in the ground. So that's the sort
of things that we do. So we've had to do hardware. We've had to do software. We have to do
networking, et cetera. And it's all powered today
on Google Cloud Platform. This is just a quick list
of some of the cities that today are using the
smart parking solution. And this is a still a
journey in progress. You know, a lot of the
things that we're doing are still just the
very basics of, just show me where
a car park is. Tell the city or the business
how the parking is used. Is it being used efficiently and
effectively and continuously, or is it underused? Or is it overused? Is the pricing right? Are the types of rules,
such as how much time you're allowed to park in a
certain parking spot, actually appropriate, et cetera? So, these are just
some of the cities that we're running today. And the big question
we're starting to pose is what about your city? So, today we already,
as you can appreciate, we've had to build a
comprehensive range of solutions. And this diagram
just illustrates some of the types of
functionality targets or audiences that this
technology delivers to. First of all, the person
in the car driving. We deliver a broad
range of applications which give guidance and
information about parking spaces. It's using Google Maps, using
Street View, all that sort of interactivity, that
convenience, that connection with you, so that you understand
what's actually happening and where are you going. And then, finally, being able
to actually pay for that parking session, whether it be
a pay-for-what-you-use-- so a timer starts to
just tick away the time, and when you've finished
your car-parking session, and you drive away, it
closes that session. We do that today. Or it can be a fixed prepaid-- you pay for 20 minutes, or pay
for 30 minutes, or whatever. So that's one class
of application that we are delivering. And just last week, we
delivered two new applications, two new versions
of that application to customers in other cities. The enforcement officer. I know this is probably
the least desirable aspect of parking. But there are people that
need to actually go along the streets and effectively
and efficiently manage where the cars that are just not
playing by the rules, right, and do the necessary
thing of putting a ticket on your window. Now, what we give is live,
real-time information to those officers on
the street telling them exactly where the car is,
exactly how much they paid, when it expire, et cetera,
et cetera, et cetera. And this is real-time
information. A lot of what Fraser
was just talking about is about being alive. The city is alive. And these sort of things
need to be constantly moving. Talking about IoT, one of
the things that we have found is a necessity of parking
is that this is actually legally enforceable IoT. As you well know,
as being drivers, and you've ridden in
cars and parked in cars-- we all have, right-- a car can actually leave a
parking space and another one pull in right away, right? And we're talking about,
wouldn't the second-- assuming you're not
speeding too fast, but it happens
really, really fast. And if we didn't detect
when that new car comes in, we could give a parking
ticket to that second arrival. We can't allow that. And this is actually the
rigor and the significance of the class of IoT
that parking is. Now, also as I
commented before, this is a battery-operated thing. And it has to last for years. But meanwhile, we're doing
this real-time communications. We've buried it in the ground,
and that's usually in concrete, or tarmac cement, and so on. It is very, very nasty
for radio communications. We've had to develop that
type of radio communications. And finally, we put
big metal things over the top of
it, which also gets in the way of doing reliable,
secure, robust radio communications. So, this is the class,
the edge of IoT, that parking is all about
and smart parking has solved. And I can tell you, we've got
a lot of wounds on our backs, a lot of scars from
learning this technology. And in fact, as I commented,
dogfooding our own technology has been a really,
really good way to make sure that we
do the right thing. What has actually started
to come back to us from our customers,
from those cities, from those customers who are
utilizing our smart parking platform, is once they put those
gateways, those communication gateways, onto their lampposts,
getting power and network connectivity to them,
either through 3G, 4G, or some sort of other
communications mechanism, that gateway becomes a very
strategic element for building out more sensing and more
capability into the smart city infrastructure and the
solutions that a city wants to actually deliver. And so, what we've been actually
being told by our customers and we're responding
to, is building an open platform,
an open framework, which enables smart cities. And so it's not
just about parking. It's actually about
modern life in cities, pretty much as Fraser and
I are talking about now. And where this is
taking us is actually into building a broad
range of solutions that sit on top of the
smart cloud platform, which is completely GCP. It's actually 100% pure
serverless as well. There is no Linux
operating system in it. There is no windows
operating system in it. It is completely built utilizing
GCP serverless computing components. And this is really
exciting because I believe this is actually
the future for how we build large-scale, internet-scale,
global-scale, live network computing systems, which
is essentially what I'm going to talk about here. Now, the first building
block that we utilized was the Cloud Core
IoT functionality that Google provided. And we did this. We started using it at
around about two years ago. Prior to that, we'd
actually been building our, own IoT framework for
managing our devices. But as we were listening
to our customers and seeing the need
for smart cities, we needed an open IoT framework. And it just did not make
sense to reinvent that wheel, especially when you had solid
serverless technology that was also going to be
driven by the market, that was going to be open and
flexible, that we didn't-- our team didn't need to rebuild. We didn't even
need to support it. We don't need to now
maintain that technology. And we can trust
that GCP and Google are going to drive
that more and more into the functionality
requirements of the market. And we just are able
to benefit from it. And as I commented
before, this is actually giving us city-scale,
live responsiveness. We are doing today with those
cities that I listed before, in the order of about a
million and a half to 2 million sensor event
transactions per day. What you might not have
noticed also on that slide, that previous one, that
diagram, was we also do cameras, especially in the UK. We do a lot of
numberplate recognition. And we're doing about a
million images per week. And so that's the
class of computing, the capacity of data
that we're processing through this platform
called SmartCloud. The following three things
that I talk about here is one of the objectives
of our SmartCloud platform was to actually
eliminate the need to have software developers
involved in everything that a customer wants. We wanted to enable
the customer to build their own dashboards,
their live dashboards, to display them anywhere
that they needed to, without requiring us to
go and hack some code. The same is true for data
analytics and data processing. And what we have
actually done is we've used BigQuery
and Data Studio, which is a fantastic tool. My analogy is Gmail to email
is Data Studio to big data. It's actually that. It's free like Gmail is. It's an amazing tool. And I'll show you in a moment
just from some displays of those sort of reports that
we're generating directly from SmartCloud, which is
using BigQuery underneath it. The same is true
for business rules. For parking, there are
very complex processes for managing business rules-- time of day when parking can
occur for certain types of uses of that parking, when is
enforcement being enabled and when is it not,
certain car parking spaces may become non-car
parking spaces at 4:30 to allow traffic flow. All those rules, including
also other things, such as, has a payment arrived? Have they overstayed their
welcome in that parking space because they've gone
beyond the maximum period you're allowed to park? So all those business
rules, which historically we wrote custom code
into the system and had some sort of a UI
that you type in some data typically into a
database, now what we have is if-this, then-that
rules engine, which processes
this continuously. And I mentioned our mobile apps
for guidance, and payments, and other things. In fact, the app area
is just so significant. We're all users of apps. What we also realized
was that you just don't want to have yet another app. There are a number of
surveys that have already shown that after 90 days or
so of getting a smartphone, you don't typically
download another app. And so getting that
footprint onto your phone is really, really hard. And so what we've built
is an app framework. Obviously, we have
full functional apps. The smart parking app on the
all stores today is our app. And a number of cities,
a number of the cities that I listed there,
have taken that app, and we've actually branded
it and facilitated it with functions that is
appropriate for them. But more interestingly is
taking those components, such as the guidance component
or the payment component, and embedding it
in your own app. And so that's the
framework for mobile apps that we've also developed. And finally, managed APIs. APIs is actually
how you integrate, how you enable
innovation with ourselves internally, how you work with
partners and software vendors, as well as integration
with city systems. And so, a full suite of managed
APIs, which are-- and in fact, is using the Apogee
technology from Google Cloud. Guess what? Just want to quickly
now show you, if we can switch to the demo-- yay. I must admit we had a little
bit of a network issue as I sort of walked up. But here's a live picture
of one of the cities. This is actually
Wellington City, the capital of New Zealand. So one end of the world. And what we've got here is,
I'm zooming in, into the city. These are the
parking bays that are around the city of Wellington. And I can see that that's
been vacant for 38 minutes. This is live data
right from the city. And this is just one
aspect of visualization. And because it's
built into dashboards, I can do such things like
push it into full screen mode and display. You can put it onto
big screens elsewhere. Every one of these
dashboards is just a URL. And one of the
powerful things is I can create my own dashboards. And there are some default ones. There's the data. SmartZone is what we
call our infrastructure area, where we can monitor
and see the infrastructure. And what you saw early
on in Fraser's slide is that sort of map of all
of the cities' activities going on. Well, here what we've got on
of the other what we call tiles is actually a 24-hour cycle
of the heat map of the parking utilization in the city. And this is just
another bit of insight that we're providing cities. Now, if you notice,
some of the things that prove this is real time, you can
see the numbers changing there. Those are parking
events that are happening on the other side
of the world right now. Here's another city. This is Hamilton City
in New Zealand as well. And I think I've
lost the other ones. I was going to also show a
couple of cities like Cardiff, and so on. So can we go back to our slides? So, just going back to
those sort of smart city requirements, one of the
things that we in Smart Parking and in the smart
parking solution have to respond to, and pretty
much what Fraser was talking about, is this live streaming,
this constant-- this instantaneous world
that we live in. And this is a change for how
computing needs to be done. It's no longer store
something to do a database, then retrieve it, do some
sort of processing on it, produce a report, and
then store it back into a database, et cetera. Things have to be
instantaneous and responsive. For example, if you were
driving down the street and you were using Google
Maps, and it took 20 seconds to sort of give
you the next turn, you'll probably be
on the wrong street up a tree or
something like that. So, this is the new
norm is live streaming. The efficiency and convenience
I've been talking just indicated the things like
apps that we are delivering the dashboards, providing those
dashboards to any display, whether it be a display on a
street, because it's a URL, anywhere there's a browser
that can be displayed. And one of the experiments that
we're actually working through is just using Android,
Android Things, to display that in remote
places onto other displays. And this is really
great for providing that convenience of information
where it's pertinent. And a dashboard can
be just, for example, how many bays are in
this area of a street, or on the floor of
a parking building? And so this is all about
connecting insights to you. And the you I mean
here is, there are many different audiences. It can be you, the
user of parking. But it may also be
the city strategist who wants to have understanding
of what's actually really going on. Are the uses and
mechanisms that are being operated for the parking,
is it optimal for the city? And as we all know, it becomes
very, very inconvenient and challenging when you're
circling round and round blocks just trying to find that park. Now using this real-time
data, accurate data, we could take you
right to that park. And when it changes, we're
letting you know as well. What we're doing is we're
starting to create the ability to enable new outcomes. And we've done that
both on the technology-- so rather than using classical
computing, and I actually class using standard
traditional servers and server OSs as
being classical, we have gone
completely serverless, completely what I call cloud
native for this platform. It is essentially the same way
that Gmail works, the same way that Google search
works, is how we've built SmartCloud, our platform. And it's enabling ourselves
and also customers to actually achieve
new outcomes. And this is driving us
to become more responsive with our roadmap. Our roadmap is no longer
a 12-month roadmap. It's actually dynamic and
changing because the way we're implementing
the functionality. And the way we're enabling
our customers to also have their roadmaps of
functionality is in a much more agile, much
more modular and free, democratized manner. Democratized is an
overused word, I think, in this conference. So here's just a quick view of
the architecture of SmartCloud. And as you can see,
it's heavily GCP. And as I mentioned,
totally serverless. So IoT Core is where we ingest
and we send out messages from and we also manage these
tens and tens of thousands of devices, both the
gateways and the sensors that are actually in the streets
or in the parking buildings. I don't know if you've
been to parking buildings where you see the red green
lights over parking bays. It's quite common in
some parts of the world, certainly is very
popular in Australasia. And it changes
the whole dynamic. You drive into your
parking building, and you literally
will just glance out of the periphery of your
vision to see a parking spot. And then we put signs
throughout just saying, OK, there are five
up on level four, and there are 12 up on
level five, and so on. And also just being able
to glance out of the side. So this sort of
information is all being managed by IoT Core,
both the measurement and also the control of those
indicators, for example. We're also managing the
data in very dynamic ways. I talked about using Big
Query as the fundamental core for big data. But simultaneously, we're using
Big Table, which is essentially the engine underneath
Google's search, to also manage those
parking events in real time. And then the API, as I
commented about, for dashboards, for payments, and
actually any application. So, over to you now. Hopefully, that's giving
you a taste of what we do. The point is that
what we're doing is real world IoT
that is trying to and is actually reinventing the
experience for cities already. [MUSIC PLAYING]
