Hi, thanks for tuning into Singularity
Prosperity. This video is the tenth in a
multi-part series discussing computing.
In this video, we'll be discussing what
cloud computing is and the fundamental
change it brings in how we view and
think about computing. To better
understand the massive revolution cloud
computing will bring and is bringing to
the field of computing, let's first go
back in time and view a similar
revolution with electricity. Before the
height of the industrial revolution,
electricity had to be generated in
house, this had many downsides: every
worker lost to generating electricity
was one less to make the factory more
productive and scalability was a major
issue, at times when production of the
factory went up, there wouldn't be enough
generated electricity, causing power
outages and a loss of production. Thus,
often times more electricity that needed
was generated which was quite costly.
Then in the 1880s, Thomas Edison, founded
the Edison Illuminating Company, turning
electricity into a utility. In other
words, something that could be switched
on and off whenever desired, delivering
the exact amount of power needed at a
cost per unit, in the case of electricity,
watts. Electricity as utility allowed for
mass productivity increases and sparked
further exponential growth in the
industrial revolution, referred to as the
second industrial revolution, this
because electricity as utility lowered the
barrier to starting and maintaining a
business, thus leading to more innovation.
Coming back to present day, this analogy
has strong correlations to
transformations seen in computing. When
running a website or application in the pre
cloud computing days, every individual
business with an online presence had to
maintain servers that allowed users to
access their site, this is referred to as
hosting. Like with electricity,
sometimes more users will access the
site and sometimes there will be little to
no users on the site. To prevent site
crashes on periods of high traffic which
therefore equated to lost users and
customers, more servers than needed often
had to be purchased. These servers are
very costly, racking up bills even when
they aren't used to full capacity. Also
like the electricity analogy, having a
large team of sysadmins, network
engineers, etc takes away productivity
from the true goal a business is trying
to achieve, therefore making the barrier to
a scalable business high and costly. Now, with
cloud computing, we are witnessing a
revolution in how computing power is
allocated, in other words, viewing
computing as a utility. Cloud computing
has seen an incremental evolution over the
past decade, in large part due to the
exponential increase in computing
performance, with cloud computing
currently growing at a rate of 23
percent per year. Before discussing the
primary types of cloud computing, what
exactly is it? Well, the best way to think
about it is to think about actual clouds.
A cloud is formed via a dense cluster of
water molecules that appear as a single
object from a distance, thus taking this
concept, cloud computing refers to a
dense cluster of computers working
together that appear as a single
computing resource. There are many
companies in the cloud computing race now,
to list some of the biggest cloud
providers: Amazon Web Services, Microsoft
Azure, Google Cloud, IBM Cloud - the list
can go on and on. Essentially, the cloud
services these companies provide are
through vast data centers made for
public use. As discussed previously, for a
business to manage its online presence,
in-house servers and maintenance was
required, which racked up costly bills
and was counterproductive to the true
goals of the business. This is where the
first type of cloud computing came into
play: Infrastructure-as-a-Service, IaaS,
where the hardware, in other words, the
hosting environment was abstracted away.
Like with an electricity meter,
businesses only pay the cloud provider
for the exact amount of computing power
used. So, when there is a heavy load on a
cloud hosted site, more is charged due to
increased computing demand, and with less
traffic, significantly less is charged.
The next type of cloud computing is,
Platform-as-a-Service, PaaS, this is where
the operating system and software
backend is abstracted away. While IaaS
provides the infrastructure for hosting
an application, everything else involved
in backend development is not covered.
This is the role PaaS fills, backend
services, including data management in
the form of databases and middleware
which is the plumbing between the
components of an application to make
sure everything works together. The last
type of cloud computing we'll discuss is,
Software-as-a-Service, SaaS, this is where
the software runtime is abstracted away,
essentially a layer in the cloud for
program execution. This part of the cloud
affects us, the consumers, the most,
allowing our devices to do minimal
processing when running an application
because processing is instead done in
the cloud and results deliver to our
devices. The best analogy for these 3
layers is a cake: IaaS is the large base of a
cake, PaaS is the frosting that covers the cake
and SaaS is the eye-catching candle on
top that people see first. Over the years,
the boundaries between these layers is
becoming more and more muddled, with the
industry heading towards a serverless
future, also called, Functions-as-a-
Service, FaaS. This is the combination of
all three types of cloud computing we've
discussed, more on this in the next
section. A serverless future provided
through cloud computing is the new
paradigm, this further exemplified by the
fact that the price of cloud computing
is decreasing due to increasing computing
power. This is referred to as Bezo's Law,
where the CEO and founder of Amazon
stated: a unit of computing power price
is reduced by approximately 50% every
3 years. The unit of measurement for
the allocation of computing power varies
by provider due to the types of cloud
computing they provide, for example, per
gigabyte of RAM used, per gigabyte of
storage, kilowatt hours of computing used,
etc. The reason for the drop in price and
increase in power for these units
extends to topics we discussed in
previous videos in this computing series
that deal with advances in both hardware
and software such as GPUs like Volta, new
memory devices as well as standards and
much more!
[Music]
When electricity became a utility, the
barrier to entry of a scalable business
dropped and led to innovation at an
exponential rate, computing as utility
does this once again. This is easily
observable by startup culture today,
where anyone with a vision or idea can
immediately establish an online presence
and proof of concept of their
application or website. Whereas in the
past, expensive servers would have to be
maintained and overhead of extra staff,
now with cloud computing and serverless
architecture all of that is taken care
of at a bare minimal cost. As the
application starts to become more
popular, all you have to do now is
increase your cloud plan through your
cloud service provider. There are many
services offered by cloud providers to
accelerate development and innovation, to
list a few: Amazon Lumberyard for game
development, Salesforce for e-commerce
applications, Amazon Web Services for all
encompassing serverless architecture and
many more. As a side note, our site
singularityprosperity.com is hosted on
AWS. Amazon Web Services offers
comprehensive big data services that
enable retailers to perform
sophisticated analytics at
lightning-fast speeds,
allowing for real-time analysis and
insights that drive your business. With
Amazon Kinesis, you can quickly ingest
different data streams from mobile and
web purchases, POS transactions and
customer data. Store this data in an
Amazon S3 data lake and use Amazon
Athena to run ad-hoc analysis and
interactive queries like customer
shopping patterns and product
preferences. With AWS Lambda you can set
up code that responds to certain
triggers and events, like querying Amazon
DynamoDB,
a NoSQL database, to respond to
customer questions about product
assortment and availability or to
generate sentiment analysis from social
media. Storing it in a fully managed data
warehouse like Amazon Redshift and then
visualizing an identifying customer
sentiment over time with Amazon Quicksight.
Amazon EMR provides retailers with
a managed Hadoop analytics framework
that makes it easy fast and cost
effective to process vast amounts of
transaction data, online browsing
behavior, in-store shopping trends and
product preferences which can guide
inventory and pricing strategies. With
the broadest and most advanced set of
analytic services supporting every
workload imaginable, AWS provides
everything a retailer needs to collect,
store, process, analyze and visualize big
data in the cloud. Beyond the impact to
business and startups, cloud computing is
beginning to have a major impact on us,
the consumers as well due to cloud
services and service architecture. Cloud
services for consumers used to just be
about storage such as Dropbox and Google
Drive.
Now, however, they are beginning to expand
to every industry, to list some examples:
cloud music like Apple Music and Spotify
and cloud gaming that Nvidia is working
on with GeForce Now. In fact, these
services are now extending towards
virtual desktops as well, where any
application can be run in the cloud. An
example of this is Amazon Workspaces,
where any Windows application can be run
in the cloud. From intensive programs
like After Effects and Cinema 4D to
productivity applications like Microsoft
Office and everything in between. In
terms of the impact of serverless
architecture on consumers, it essentially
means our devices will be able to do
more with less processing power, since
the most computationally intensive tasks
will be computed in the cloud. Let me
explain this and extrapolate the future
of cloud computing. There are three
layers: cloud data centers, fog nodes and
edge devices. Cloud data centers are the
giant facilities that companies such as
Amazon and Google have, the scale of them
globally are in the thousands. Fog
nodes are small-scale data centers or
servers that can be installed anywhere,
from the floor of a building, inside a
railcar - the list can go on and on. Simply
put, any device with computing, storage
and networking is a fog node and their
scale will be in the millions. Edge
devices are simply any
internet-connected device, from our
mobile phones, self-driving cars, sensors
and the countless IoT devices that are
coming to market, their scale will be in
the billions. With the growth of machine
learning and artificial intelligence,
more and more edge devices will require
instantaneous big data insights,
such an example is self-driving cars
that are constantly learning and
adapting to their environments. It is
impractical to have edge devices
themselves be powerful enough to compute
these analytics, but what they can do is
make inferences based on current machine
learning models they have and acquire
large amounts of data. This newly
acquired data is then sent to the cloud
so updated machine learning models can
be made and then sent back to the edge
device.
The purpose of fog nodes is to reduce
the latency of access to the cloud due
to the vast amounts of devices that will
need connectivity. Instead of connecting
to a data center far away, the fog acts
as an intermediary that stores analytics
closer to the edge. We will cover edge
computing much more in depth in this
channels AI and Internet of Things
series! Also if you're curious about
issues of latency between the edge, fog
and cloud - be sure to check out this
channels previous series on 5G and video
on optical computing! At this point the
video has come to a conclusion, I'd like
to thank you for taking the time to
watch it. If you enjoyed it, consider
supporting me on Patreon to keep this
channel growing and if you want me to
elaborate on any of the topics discussed
or have any topic suggestions, please
leave them in the comments below.
Consider subscribing for more content,
follow my Medium publication for
accompanying blogs and like my Facebook
page for more bite-sized chunks of
content. This has been Ankur, you've been
watching Singularity Prosperity and I'll
see you again soon!
[Music]
For more infomation >> Our King is Marching On - Duration: 29:13. 
For more infomation >> What is marriage? Franche Primary School Royal Wedding - Duration: 2:24. 
For more infomation >> MLMA UN-CUT/ 03. Who said Modeling is easy - Duration: 9:22. 
For more infomation >> The Last Enemy to be Destroyed is Death - Duration: 29:26. 
For more infomation >> Introducing Tick Talks | This is That - Duration: 3:02. 
For more infomation >> Peter Schrager: The sky is the limit for the 49ers in 2018 - Duration: 1:55. 
Không có nhận xét nào:
Đăng nhận xét