►
From YouTube: Introduction to Cloud
Description
An introduction to what the cloud actually is and its benefits.
🔎 Looking for content on a particular topic? Search the channel. If I have something it will be there!
▬▬▬▬▬▬ C H A P T E R S ⏰ ▬▬▬▬▬▬
00:00 - Introduction
01:11 - Layers of your computer
04:50 - Datacenters
10:32 - Cloud regions
16:30 - Cloud capacity and resource
19:35 - Compute services
21:42 - Moving responsibility
24:12 - Infrastructure as a Service
26:06 - What is most important
26:44 - Platform as a Service
28:38 - Software as a Service
31:13 - Other types of resource offered
33:58 - Identity
37:19 - Benefits of the cloud
38:05 - Multi-tenant
39:26 - Broad network access
43:38 - Provision on-demand
46:12 - Measured service
49:03 - Self-service
51:12 - Summary
▬▬▬▬▬▬ K E Y L I N K S 🔗 ▬▬▬▬▬▬
► Whiteboard:
🔗 https://github.com/johnthebrit/RandomStuff/raw/master/Whiteboards/IntroductiontoCloud.png
► Azure regions:
🔗 https://datacenters.microsoft.com/globe/explore
► Latency examples:
🔗 https://learn.microsoft.com/azure/networking/azure-network-latency?tabs=Americas%2CEastUS
▬▬▬▬▬▬ Want to learn more? 🚀 ▬▬▬▬▬▬
📖 Recommended Learning Path for Azure
🔗 https://learn.onboardtoazure.com
🥇 Certification Content Repository
🔗 https://github.com/johnthebrit/CertificationMaterials
📅 Weekly Azure Update
🔗 https://youtube.com/playlist?list=PLlVtbbG169nEv7jSfOVmQGRp9wAoAM0Ks
☁ Azure Master Class
🔗 https://youtube.com/playlist?list=PLlVtbbG169nGccbp8VSpAozu3w9xSQJoY
⚙ DevOps Master Class
🔗 https://youtube.com/playlist?list=PLlVtbbG169nFr8RzQ4GIxUEznpNR53ERq
💻 PowerShell Master Class
🔗 https://youtube.com/playlist?list=PLlVtbbG169nFq_hR7FcMYg32xsSAObuq8
🎓 Certification Cram Videos
🔗 https://youtube.com/playlist?list=PLlVtbbG169nHz2qfLvPsAz9CnnXofhmcA
🧠 Mentoring Content
🔗 https://youtube.com/playlist?list=PLlVtbbG169nGHxNkSWB0PjzZHwZ0BkXZZ
❔ Questions? Maybe I answered it in my FAQ
🔗 https://savilltech.com/faq
👕 Cure Childhood Cancer Charity T-Shirt Channel Store
🔗 https://johns-t-shirts-store.creator-spring.com/
👂 Enable the subtitles and from there you can translate to your native language via the auto-translate feature in settings! https://youtu.be/v5b53-PgEmI for a demo of using this feature.
SUBSCRIBE ✅ https://www.youtube.com/channel/UCpIn7ox7j7bH_OFj7tYouOQ?sub_confirmation=1
#microsoft #azure #johnsavillstechnicaltraining #onboardtoazure #cloud
A
A
In
this
video
I
wanted
to
provide
an
introduction
to
the
cloud.
We
always
talk
about
the
cloud,
and
so
as
I
was
talking
to
my
kids,
the
other
day,
I've
really
tried
to
go
through
and
explain
some
of
the
basics
about
it.
So
I
thought
I'd
record
it
as
well.
Maybe
try
and
provide
some
help
for
other
people
who
are
maybe
new
to
technology
or
maybe
shifting
their
focus
to
start
understanding
and
using
some
of
these
cloud
services.
A
Now
I
do
want
to
stress
when
I'm
going
to
talk
about
the
cloud
I'm
going
to
focus
on
the
public
Cloud,
which
is
what
most
people
think
of
when
they
say
the
cloud
now.
Obviously,
there
are
ways
that
I
can
have
a
private
Cloud
within
my
company,
where
I
take
my
servers
and
I
have
a
special
type
of
management
infrastructure
that
exposes
some
of
the
capabilities.
I
very
commonly
in
the
company
will
have
hybrid,
where
I
have
some
public
clouds.
A
Some
private
there's
Community
clouds
that
maybe
certain
agencies
gather
together
and
share
some
infrastructure,
but
we're
going
to
focus
on
the
public
Cloud,
which
is
what
most
people
really
think
about
when
they
say
the
word
cloud,
but
I
actually
want
to
take
a
step
back
for
a
second.
So
before
we
start
talking
about
the
public
cloud
and
everything
that
entails
so
I
think
we
can
all
probably
relate
to
and
what
we're
used
to
is
the
idea
of
well
our
computer.
A
There
were
different
aspects
that
I
have
to
think
about
now.
My
computer
I
buy
it
and
it
has
certain
amounts
of
capacity
that
I
can
use,
so
I
can
think
about.
Well,
it
has
a
certain
CPU,
maybe
different
Generations
different
vendors
and
I
have
a
certain
amount
of
Random
Access
Memory
that
can
be
used
by
the
applications
that
are
running
so
I
have
a
certain
amount
of
compute
I
have
storage
to
store
my
data
to
install
my
applications
on
this
might
be
a
hard
disk
drive
more.
A
I
have
certain
types
of
network
now
that
Network
May
provide
connectivity
between
things
within
my
home
or
at
my
office,
and
then
there's
connectivity
from
my
location
out
to
the
internet
and
maybe
through
that
I
also
have
maybe
some
private
connections
to
my
corporate
resources,
and
then
we
install
an
operating
system.
Now.
A
All
of
this
remember
is
capacity
that
I
can
do
things
with
and
then
I
install
some
operating
system
or
maybe
come
pre-installed
Windows
Linux
Mac
OS
whatever
that
might
be,
but
what
this
capacity
is
providing
to
you
in
this
case
is
an
operating
system.
That's
your
resource!
Now,
on
top
of
that,
you
install
your
stuff,
the
stuff
you.
A
I
installed
my
applications,
this
could
be
productivity,
apps
like
an
office
application.
It
might
be
games
that
I
want
to
play
whole
different
set
of
things
web
browsers
and
then
I
have
my
data.
The
documents
I
create,
maybe
media
I
leverage
whatever
that
might
be,
and
then
to
use
my
machine
and
when
I
interact
with
different
things.
I
have
identities.
A
To
log
on
to
my
machine,
maybe
that's
a
local
identity,
maybe,
when
I
go
and
interact
with
certain
resources,
there's
a
corporate
identity
when
I
go
and
use
cloud-based
applications.
Well,
maybe
there's
a
cloud
based
identity
provider.
Now
that
could
be
someone
like
a
Microsoft
account.
It
could
be
a
Facebook
account,
a
Google
account
and
often
you'll,
see
that
hey
I
can
take
that
cloud
identity
and
use
it
in
many
different
places.
A
I
don't
have
to
have
an
identity
for
every
single
application.
It's
like
hey
I,
want
to
go
and
use
this
site.
It
says:
hey.
Do
you
want
to
sign
in
with
an
existing
provider
a
social
provider
identity?
You
already
have
so
I
get
this
idea
of
a
single
sign-on
I,
don't
have
to
have
an
account
for
every
single
company
which
can
be
fairly
difficult
to
maintain.
C
A
A
So
we
have
this
idea
of
now.
They
may
have
multiple,
but
we
have
these
data
centers,
and
it's
really
not
that
much
different
from
your
PC.
These
data
centers
are
filled
up
with
these
big
racks.
These
structures
that
have
a
lot
of
power
going
into
them
and
network
switches
that
I
install
really
computers
into
them,
and
so
what
these
data
centers
are
providing
to
me
is
that
same
idea
of
capacity.
A
It
provides
me
capacity
because
they
have
that
same
idea
of
networking.
They
have
storage
systems,
they
have
lots
of
servers
providing
me
compute,
but
it's
fundamentally
given
me
that
amount
of
capacity
now
one
thing
that
definitely
will
be
different
in
any
all
data
centers.
We
drew
the
idea
here
of
the
PC
and
then
directly
on
the
hardware
directly
onto
them,
the
motherboard
and
the
storage
we
had.
We
installed
the
operating
system,
so
there
was
a
very
definite
connection.
A
A
Maybe
have
different
operating
system
versions
and
flavors,
hey
I've,
Linux
I
have
Windows,
I
have
Ubuntu,
I
have
sentos,
I
have
susei,
I,
have
rocky
looks
whatever
that
might
be,
they
run
different
libraries
different
Frameworks
and
I
want
to
be
able
to
take
these
and
maybe
move
them
around
different
physical
bits
of
hardware
and
not
worry
about
or
which
driver
do
I
particularly
need.
So
what
you're
seeing
is
in
these
data
centers?
A
A
And
what
the
hypervisor
does
is
it
provides
this
abstraction
between
the
physical
Hardware
and
then
it
creates
this
virtualized
set
of
Hardware,
so
the
whole
point
of
a
hypervisor.
Is
it
virtualizes
the
hardware
it
says
hypervisor
lets
me
create
virtual
machines
that
have
their
own
allocations
of
a
certain
amount
of
virtual
CPU
or
compute.
They
can
have
a
certain
amount
of
memory
assigned
I
can
expose
different
amounts
of
storage
to
them,
which
then
go
and
get
created
on
the
underlying
storage.
A
B
A
Can
then
install
well
in
this
one
I'm,
maybe
going
to
install
a
version
of
Windows
hey
in
this
one
I'm
going
to
install
a
version
of
Linux
in
this
one?
It's
a
different
version
of
Linux.
This
requires
this
run
time
for
this
app
to
work.
This
has
a
different
set
of
libraries.
It
wants
to
use
for
a
different
application,
etc,
etc,
and
the
key
Point
what
this
hypervisor
is
providing
is
very,
very
strong
control
of
the
amount
of
resource,
but
also
isolation.
A
Just
because
they're
running
on
the
same
physical
piece
of
equipment,
this
hypervisor
guarantees,
hey
I,
can't
go
and
see
each
other's
processes,
so
this
sits
between
the
physical
hardware
and
then
these
virtual
machines,
but
it's
still
just
capacity
that
I'm
exposing
as
a
certain
type
of
resource
and
one
of
the
nice
things.
Obviously,
we've
used.
These
virtual
machines
is
I,
get
better
use
of
the
physical
equipment.
I
can
create
these
virtual
machines
near
instantly.
I'm,
not
waiting
for
someone
to
say,
hey
I
need
a
new
OS
instance
to
run
my
business
app.
A
Please
go
and
order.
This
server
wait
for
it
to
arrive,
rack
the
server
cable.
The
thing
up
and
then
install
the
OS,
hey
I,
need
a
new
VM.
It
can
be
up
and
running
in
minutes,
so
this
is
all
just
really
boiling
down
to
this
idea
that
we're
used
to
the
idea
of
having
capacity
that
we
expose
as
some
type
of
resource.
That's
really,
all
any
of
this
is
doing
it's
giving
me
these
compute
the
storage,
this
network,
that
I
then
expose
as
some
other
type
of
resource,
really
nothing
more
than
that.
A
Then
we
have
this
idea
of
the
cloud
again
and
there's
jokes:
hey
the
cloud
is
just
someone
else's
data
center,
there's
no
such
thing
as
the
cloud
and
there's
an
element
of
Truth
to
that
there
is
no
magic
vaporware
running
this
thing.
There
are
lots
and
lots
of
servers
in
lots
and
lots
of
locations
that
actually
make
this
thing
work
and
what
you'll
see
for
all
of
these
providers.
Now
this
could
be
Azure.
This
could
be
AWS.
This
could
be
gcp.
A
There
are
many
many
others
of
these,
but
they're
all
going
to
have
this
idea
of
regions
and
a
region
would
be
a
certain
area
of
space
on
which
they
put
multiple
data,
centers
and
then
I
have
these
all
around
the
world.
So,
if
I
think
about
okay
I'm
this
cloud
provider
here,
whichever
one
I
am
hey,
they'll
see
an
idea
of
a
region,
one
and
I'll
see
a
region
two
and
a
region,
three,
etc,
etc,
and
then,
even
within
those
regions,
it
won't
just
be
the
idea
of
one
of
those.
A
A
So,
if
I'm
thinking
about
deploying
my
application,
that
really
is
what
I
care
about
well
I
can
make
sure
it's
distributed
over
different
data
centers
through
the
data
center
caught
fire
or
had
some
kind
of
power
substation
failure.
My
app
would
still
keep
running
even
within
that
region.
Providing
I
have
it
running
in
multiple
zones
within
that
particular
region,
and
we
can
see
examples
of
this.
A
So
if
I
was
to
just
jump
over
and
we
just
look
at
one
example,
if
I
look
at
Azure,
because
that's
what
I
use
all
the
time
so
Azure
these
blue
dots
are
existing
regions
and
I
can
go
and
look
around
the
world.
In
fact,
let's
flatten
this
out
and
I
can
see.
Well
there
there's
lots
of
regions
throughout
the
United
States.
For
example,
West
U.S
three
one
there'd
be
a
two
there's
one's
in
Canada
I've
got
ones
down
here
in
Brazil
locks
in
Europe,
UAE,
India,.
A
We
have
these
regions
all
throughout
the
world
and
again
it's
not
one
data
center.
There's
lots
of
data
centers
within
those
particular
regions.
Now
you
may
wonder
why
bother?
Why
do
I
need
more
than
one
location
and
it's
all
about,
we
always
think
of
light
as
really
really
fast
and
instant,
but
it's
not
light
takes
time
to
get
from
one
place
to
the
other
place.
It's
not
instant.
It's
actually
slow
in
terms
of
thinking
about
computers.
A
If
I'm
in
my
home
and
I'm
talking
between
my
different
devices,
it's
instant,
they
can
be
as
chatty
and
going
back
and
forth
as
they
want
it
won't
matter,
but
if
I
try
and
work
with
a
computer
on
the
other
side
of
the
world.
Well,
let's
think
about
that.
So
if
I
jump
back
over
for
a
second
here
are
some
example
latencies.
So
latency
is
just
the
time
it
takes
to
send
some
packet
of
information.
A
A
It's
like
well
I'm,
really
not
worried
about
a
fifth
of
a
second
to
do
anything,
I'm
not
going
to
notice,
but
if
you
have
computer
systems
that
are
frequently
talking
like
an
application
to
its
database
or
some
replication
fifth
of
a
second
is
a
really
long
time.
There
doesn't
have
to
be
many
interactions
to
now.
If
I
think
of
a
certain
operation
I
perform
is
a
hundred
interactions.
Well,
that's
a
really
really
long
overall
latency
for
the
interaction
and
it
wouldn't
work.
A
If
I
was
a
customer,
I
was
going
to
a
Commerce
site
and
they
only
hosted
in
one
place,
and
it
took
me
20
seconds
to
click.
Any
link
I
would
consider
that
amount
of
latency,
broken
and
I
would
go
away.
So
the
idea
of
having
these
multiple
regions
is
I
want
to
provision.
My
services
create
them
near
to
where
they're
going
to
be
used
so
there's
a
short
physical
distance
to
travel,
which
will
therefore
be
a
lower
latency,
and
so
that's
why
they
have
these
many
regions.
A
A
These
would
be
hundreds
of
miles
apart.
So
even
if
there
was
some
disaster
that
impacted
one
of
my
regional
deployments,
either
I'm
already
running
in
another
one
as
well,
we
call
that
active,
active
or
I
have
the
ability
to
start
my
services,
it's
passive,
but
it
can
start
up
and
then
run
my
services.
So
that's
why
we
like
this
idea
of
lots
and
lots
of
regions,
so
it
gives
the
customer
flexibility
the
customer
can
put
Services
wherever
they
may
need
to
be.
A
A
It's
not
magic,
there's
a
bunch
of
servers
running
a
bunch
of
storage
with
lots
of
connectivity,
but
it
exposes
it
as
many
different
types
of
resource.
Now
we
used
the
idea
if
we
think
back
to
that
on
premises
scenario
when
we
expose
them,
maybe
as
an
operating
system,
maybe
we
expose
them
as
a
virtual
machine.
You
may
hear
the
word
containers
and
kubernetes
kubernetes
helps
you
manage
containers
and
a
container
is
simply
if
a
virtual
machine,
virtualizes
Hardware
a
container
lets
you
virtualize
the
operating
system.
A
So
now,
even
within
the
operating
system,
I
can
further
create
these
sandboxes.
These
isolations
between
different
sets
of
processes
that
I
can
control
the
amount
of
resource
they
can
use
and
I
can
control
their
interactions.
What
they
can
see
between
each
other
and
because
now
I'm
sharing
an
operating
system,
there's
less
overhead
and
I
can
create
them
even
faster
than
even
creating
a
virtual
machine.
A
So
we
have
the
idea
of
okay.
We
we
want
to
provide
these
resources.
So
if
we
were
limited
in
what
we
could
create
up
here,
if
I
think
of
the
cloud
well,
just
the
sheer
scope
and
the
type
of
resources
they
can
create
is
far
greater.
They
have
a
lot
more
flexibility
because
of
the
sheer
scale
because
of
the
amount
of
money
they
can
spend
on
research
and
development
and
everything
else
they
can
really
expand
to
types
of
service.
That
would
not
be
practical
for
us
to
try
and
run
on
premises.
A
If
you
consider
the
recent
generative
AI
GPT,
for
example,
behind
the
scenes
there
are
thousands
of
gpus
running
that
hey
I
type
in
my
little
question,
and
you
see
the
the
words
appear,
those
tokens
thousands
of
gpus
are
powering
those
models.
I
couldn't
run
that
in
my
data
center
I
could
not
afford
to
buy
thousands
of
gpus.
A
They
also
run
at
efficiencies
in
terms
of
the
power
that
comes
in
and
the
amount
of
power
they
actually
use
for
compute
this
power
usage,
Effectiveness
Pue,
will
never
match
those
numbers
they
think
about
how
much
water
do
they
use?
What's
the
impact
on
the
environment
so
there's
a
whole
set
of
things,
they
can
do
that.
We
just
can't
match
when
we
think
of
on-premises,
but
let's
just
take
it
to
a
really
basic
one
of
the
most
common
types
of
service.
We
ever
think
about
is
compute.
A
A
And
then
remember,
we
want
to
put
a
hypervisor
on
top
of
it
and
then
inside
that
virtual
machine
we
get.
We
have
operating
systems,
maybe
one
times
libraries
middleware,
and
then
we
get
to
what
we
actually
care
about
as
a
company
which
is
our
applications,
our
business
apps
and
our
data
and,
if
I,
think
about
what
I'm
used
to
so.
If
I
think
of
my
on-premises.
A
The
responsibility
is
you
for
all
of
that,
for
every
single
aspect
of
that
now
I
might
have
different
teams
doing
it,
but
I
am
responsible
for
every
aspect.
I
need
a
storage
team,
a
networking
team,
a
server
team,
an
OS
team
I
do
all
of
that.
Now.
As
soon
as
we
move
to
the
cloud,
we
start
to
shift
this.
So
if
I
now
think
of
what
I
get
in
the
cloud,
the
most
basic
service
I
will
get
is
something
called
infrastructure
as
a
service.
A
So
infrastructure
is
a
service.
Is
a
VM
in
the
cloud.
That's
probably
the
simplest
way
to
think
about
it,
but
before
we
even
go
into
that
idea
of
that
infrastructure
as
a
service,
think
back
just
for
a
second
it's
my
idea
over
here
of
my
computer
and
remember:
we
had
this
idea
that
well
there
was
the
operating
system
and
all
of
the
hardware
then
I
could
install
the
apps
I
cared
about.
So
there
was
a
whole
bunch
of
responsibilities
that
I
had
to
do.
A
I
was
responsible
for
what
I
had
to
patch
the
computer
I
would
have
to
patch
the
operating
system.
Drivers,
I
would
have
to
patch
various
run
times
that
required
for
the
apps
I
want
to
run.
I
have
to
worry
about
things
like
backups
and
anti-virus
firewalls
config,
because
there's
a
general
purpose
system
now
I
can
do
anything,
it's
very,
very
flexible
by
the
whole
set
of
responsibilities
that
I
have
to
do.
A
Well,
then,
I
might
think
about
well
also,
what
I
have
is
a
games
console
now,
if
I
think
about
my
games,
console
I,
don't
do
any
of
this?
That
responsibility
shifted
away
from
me.
I,
don't
have
to
worry
about
any
of
those
things
for
my
games
console,
but
it's
more
specific
in
what
I
can
do
where
I
say
apps
up
here.
Really
what
I
mean
up
here
now
is
I
can
run
games.
A
So
it's
now
become
a
more
specialized
type
of
service
and
it
shifted
the
responsibility.
There's
still
an
operating
system,
they're
still,
drivers
they're
still
run
times,
they're
still
data
happening
on
that
thing.
That's
maybe
backed
up
to
the
cloud,
but
I'm
not
responsible
for
that
at
all
anymore.
The
manufacturer
of
that
games
console
is
now
providing
me
a
different
type
of
equipment
and
they
take
ownership
of
all
those
responsibilities
that
are
still
there,
but
I
now
only
worry
about
really
what
I
care
about.
A
In
this
case,
yes,
I
can
run
games
on
a
PC,
and
many
of
us
do
at
bike
and
one
other
stuff
as
well,
but
I
take
on
a
bunch
of
responsibility
with
a
games.
Console
I,
don't
worry
about
what
is
this
console
good
enough
to
run?
This
game
is
The
Game's
available
for
it.
It's
been
tested
specific
on
that
I
know
it's
gonna
work,
I've
shifted
responsibility,
the
things
I
actually
don't
care
about,
but
I
lost
a
bit
of
flexibility.
A
A
We're
shifting
responsibility
all
of
the
physical
aspect
of
the
compute,
the
servers,
the
network,
the
hypervisor
that
is
now
the
responsibility
of
the
cloud
provider
I
had
no
access
to
the
direct
physical
equipment,
I
wouldn't
even
know
what
box
I'm
on
what
I
am
now
responsible
is
for
everything
in
this
case
inside
that
virtual
machine.
Now
what
that
means
is
I
can
pick
the
operating
system.
Hey
I,
want
to
install
a
certain
Linux
distribution.
A
I
want
to
install
a
certain,
maybe
Windows
version,
but
it
also
means
I
have
now
shifted
or
worry
of
responsibility
about
the
physical
equipment,
but
I'm
still
responsible
for
things
inside
of
that
virtual
machine.
So,
for
example,
I
would
still
be
responsible
for
thinking
about
well
patching
I'd
be
responsible
for
thinking
about
backing
it
up
for
antivirus.
I
would
think
about.
Dr
do
I
need
to
do
replication.
A
C
A
A
huge
amount
of
flexibility,
it's
a
VM,
I'm
picking
the
operating
system.
I
can
run
anything
I
want
inside
that
virtual
machine,
providing
there's
not
some
weird
quirky
Network
requirements
that
don't
work
in
a
public
cloud
model,
so
fantastic
and
the
most
basic
level
I
could
just
create
a
VM
in
the
cloud
and
I
can
use
those
remember
that
idea
of
Shifting
responsibility
with
a
VM
in
the
cloud.
Yes,
I've
shifted
responsibility
for
the
physical
stuff
and
the
hypervisor
I
have
a
lot
of
responsibility
and
remember
for
most
organizations.
A
They
don't
really
want
to
be
in
the
business
of
managing
any
of
that
stuff.
What
they
really
care
about
is
their
data
and
their
business
applications
that
help
differentiate
them
from
any
other
company.
So,
if
possible,
they
want
to
offload
as
much
stuff
that
isn't,
differentiating
them
and
generating
revenue
from
someone
else.
A
A
A
It's
not
my
responsibility,
anymore,
I'm,
Shifting,
the
responsibility
to
the
cloud
provider
and
I'm
only
responsible
now
for
my
custom
business
application
and
the
data
associated
with
it
I've
offloaded
that
and
what
we'll
see
here
is
there's
a
lot
of
different
types
of
service
available.
You
will
hear
obviously
about
containers
and
kubernetes
is
the
standard
in
orchestration.
Either
management,
the
creation,
the
health
checking
of
those
container
environments,
you'll
have
web-based
applications,
there'll
be
Services,
particularly
for
web
applications.
There'll
be
batch
processes.
A
You'll
hear
about
serverless
serverless
enables
me
to
run
a
bit
of
code
or
a
flow
that
I
write
on
some
trigger.
It's
event,
driven
an
event
happens
that
event
could
be
a
schedule
that
event
could
be
it's
manually
triggered
that
event
could
be.
Some
signal
comes
in
through
an
API
or
something
happens
on
another
service
that
makes
it
run
that
bit
of
code,
but
I'm,
shifting
more
and
more
responsibility
to
the
cloud
provider.
A
A
Now
these
will
be
more
specialized.
These
might
be,
for
example,
productivity
applications,
things
like
Microsoft
365.
It
could
be
a
CRM
system,
it
could
be
a
HR
System,
there's
a
huge
set
of
these,
but
this
will
all
be
the
responsibility
of
the
provider
of
the
solution.
You
don't
you're,
not
responsible
for
anything.
You
might
do
basic
Administration
to
enable
people
for
certain
roles
or
features,
but
you
are
not
upgrading
exchange
online
I
am
not
patching.
Salesforce
I
am
not
backing
up
my
own
data
in
these
things,
it's
done
as
part
of
the
service.
A
So
all
of
these
again
it's
all
about
picking.
Well,
what
is
it
I
really
really
care
about?
What
is
it?
I
need
to
achieve
how
much
responsibility
can
I
shift
so
that
I
don't
have
to
worry
about
it.
People
make
the
same
decision
here.
Hey
my
goal
is
I
just
want
to
play
a
game.
I,
don't
want
to
worry
about
managing
these
things,
I
buy
a
games
console,
but
maybe
actually
I
need
some
more
flexibility.
I
want
to
be
able
to
run
productivity
apps.
A
Well
then,
this
is
too
specialized,
and
so
there's
going
to
be
times
that,
although,
if
we
can,
we
want
to
go
to
the
SAS
route,
we
want
to
be
responsible
for
nothing.
Well,
maybe
I'm!
Writing
my
own
custom
business
application
that
differentiates
me
from
my
competitors.
There
isn't
a
SAS
solution,
so
I'll
create
a
pad
solution
and
if
I
can
do
serverless
I
will
maybe
it's
web-based.
A
Maybe
it's
container
based,
but
maybe
there
will
be
times
where
maybe
it's
a
legacy
app
that
I'm
moving
into
the
cloud
or
I
just
have
some
requirements
that
requires
me
to
interact
directly
with
the
operating
system.
I
have
very
stringent
requirements
around
the
runtime
of
libraries.
Well,
then,
I
would
pick
I
as
you
have
the
flexibility.
It's
your
choice,
which
one
makes
the
most
sense
for
you
based
on
your
requirements.
I
mean
that
that's
the
whole
point
about
all
of
these
different
things.
It's
your
choice
to
use
the
one
that
you
need
to
do
now.
A
A
Ability
to
create
basic
things
like
blobs
that
can
store
anything
unstructured
data,
there'll,
be
queues
and
tables;
there'll
be
file.
Based
Services,
like
SMB
and
NFS
there'll,
be
caching
services
like
redis
there'll,
be
databases,
databases
that
may
be
built
on
open
source
Technologies
like
postgres
or
MySQL
there'll,
be
ones
built
on
without
Microsoft
SQL
Server
they'll
be
cloud
based:
multi-master
Services
things
like
Cosmos
DB
in
Azure,
for
example,
document-based
databases,
there'll,
be
analytic
services.
A
I'll
have
search
Services
text,
space
Vector
based
so
we'll
hear
vectors
all
the
time
now
vectors
enable
me
to
run
images
and
audio
and
text
through
a
model
to
create
this
numeric
Vector.
That
represents
that.
Well,
that
Vector
is
then
very
useful
to
be
able
to
then
search
based
on
whatever
I
want
to
type.
Maybe
it's
another
image,
but
it
runs
through
the
same
model
and
then
it
can
find
distances
to
what's
the
closest
match.
A
So
it
lets
me
now
not
have
to
be
exact
in
my
wording,
but
this
is
what
things
like
AI
make
huge
advantage
of.
So
we
have
these
analytics.
We
have
these
search
services
and,
as
I
just
mentioned,
we
have
these
artificial
intelligence,
machine
learning,
Services,
there's
things
like
Internet
of
Things
chip
in
everything.
Everything
has
an
IP
address.
Well,
they
have
to
talk
to
something
they
have
to
report
their
metrics
into
it.
A
A
A
A
Now
this
identity
provider
May
synchronize
from,
for
example,
an
on-prem-
you
may
have
an
existing
active
directory
domain
Services.
It's
very
common
I,
may
absolutely
sync
that
so
it's
still
seamless
for
the
user
into
that
cloud:
identity
provider,
but
it's
this
identity
provider
that
will
power
the
identity
used
for
all
these
services
and
if
I
have
other
cloud
services,
I
use
I,
don't
want
a
different
identity
for
all
of
those.
A
So
once
I
have
this
Cloud
identity
provider,
all
of
your
cloud
services
can
trust
it
and
then,
as
the
user
I
get
this
great
single
sign-on
experience.
It's
one
identity,
I
access
everything
with
that
one
identity:
it
is
the
new
perimeter,
often
on
premises.
We
think
about
the
network
is
the
perimeter
because
my
staff
lives
in
the
data
center.
I
have
to
have
a
network
connection
to
be
able
to
talk
to
it,
so
we
think
network
network.
A
Zero
trust
has
shown
us
that
it's
a
very
poor
idea,
because
they're
saying
infiltrates,
your
network,
you're
done
I
want
to
always
be
rechecking.
Permissions
I
want
to
always
be
looking
for
signals
that
there's
a
problem,
so
we
really
want
to
lock
down
this
identity,
which
is
why
you'll
hear
things
about
multi-factor
authentication
you'll,
hear
about
removing
the
password
completely
and
going
passwordless
you'll
hear
the
idea
about
using
policy
to
understand
what
are
they
trying
to
do?
A
I
might
get
signals
like
Risk
and
if
it
seems
like
it's
risky,
it's
a
weird
IP
I'm,
getting
multiple
logins
from
distances,
that's
physically
impossible
to
travel,
hey
I'll
apply
a
more
stringent
policy,
I'll
make
them
do
a
stronger
type
of
multi-factor
authentication
or
maybe
insist
on
also
you're
coming
from
a
certain
secured
segment.
There's
a
certain
device
has
to
be
registered
and
known
to
be
healthy.
I
can
have
all
these
different
things
come
in
as
part
of
that
identity,
which
is
powering
everything
you're
going
to
do
so.
A
I
want
the
idea
of
a
single
identity
that
I
can
highly
secure.
Highly
monitor,
look
for
any
signs
of
risk,
so
this
this
is
what
the
cloud
is.
The
cloud
is
not
some
magic,
mystical
thing,
it's
the
same.
Physical
things
we're
used
to,
but
what's
really
powerful,
is
often
on
premises.
We
end
up
with
these
islands
these
silos
of
resource,
and
so
we
have
to
buy
news
servers
for
a
particular
project
and
it
takes
time
and
it's
slow.
Someone
else
has
to
provision
things
for
me.
A
A
A
In
our
data
center,
it's
us
it's
our
company,
it's
just
me
and
the
cloud
it's
absolutely
generally
open
to
anyone.
That's
a
public
cloud
you'll
often
see
the
big
vendors
like,
for
example,
Azure
has
a
special
gov
Cloud,
where
any
government
customers
could
do
because
they
want
an
even
additional
level
of
assurance
on
who
else
they're
interacting
with,
but
it's
multi-tenant
is
very
rarely
a
physical
box
per
customer.
A
It's
this
huge
pooling
of
resources,
but
then
it's
trusting
and
it's
leveraging
technology
to
provide
that
isolation
and
control
the
hypervisor,
their
software-defined
networking.
But
fundamentally
it's
pulling
everything
together
and
then
carving
it
up
using
software
to
ensure
isolation.
There's
encryption
going
on
at
all
levels
to
give
me
that
assurance
that
well,
there's
no
risk
in
sharing
the
physical
boxes.
A
A
The
word
Public
Access
I
want
to
split
that
in
two
parts,
because
I've
used
the
word
control
and
management
a
few
times.
So,
if
I
think
about
public
access,
it
doesn't
mean
every
resource
I
create,
is
open
to
the
internet,
that'd
be
pretty
horrible.
I
I
don't
want
to
do
that,
but
what
it
means
from
Public
Access
is,
if
I
think
about
there's
a
control
plane,
so
a
control
plane
or
you
might
think
of
it
as
a
management
plane,
is
what
I
go
and
interact
with
now.
I
might
interact
with
it
through
a
graphical
portal.
A
It
might
be
through
a
command
line
interface
through
a
Powershell
through
a
template
for
a
restful
API
that
is
publicly
accessible.
Now
that
control
plane
is
secured
through
role-based
access
controls
through
policies
that
identity
provides
that
interface
in
it's
not
just
an
open
endpoint,
but
those
create
update,
delete
operations.
A
Well,
they're
done
through
this
control
plane
and
that
control
plane
yes,
is
accessible
to
the
internet.
Now
again,
I
can
lock
it
down
if
I
want
to.
But
that's
a
key
aspect
of
this
I.
Don't
have
to
have
some
special
set
of
connectivity
to
start
using
the
public
Cloud.
Anyone
could
just
go
and
get
up
and
running
with
this.
So
it's
got
broad
network
access
now,
when
I
think
about
the
data
plane.
A
Now
the
data
plane
is
what
what
is
it
actually
offering
from
within
that
resource
now
for
a
VM
I
would
think
of
that
as
the
RDP
or
an
SSH
connection
to
actually
talk
to
it
if
I
go
and
create
that
I
as
that
VM
in
the
cloud
I
don't
want
to
just
give
it
a
public
IP,
which
means
it
can
be
talked
to
from
the
internet.
If
I
do
that,
you
can
have
a
thousand
people
trying
to
hack
it
now
sure
you
might
have
a
really
good
password,
but
it's
not
a
good
idea.
A
A
So
I
could
have
certain
checks
as
you're
trying
to
come
in
and
talk
to
it.
I
can
do
things
like
just
in
time
or
just
in
time
will
do
is
it
will
only
open
up
a
gap
in
the
firewall,
for
maybe
the
IP
address,
I'm
at
and
only
for
a
limited
window
of
time,
but
the
nicer
thing
I
could
do
is
if
I
think
that
well
actually
I
also
still
have
my
physical
locations.
A
I
can
enable
a
connection.
So
I
can
have
a
connection
from
my
on-premise
location
to
the
cloud
provider
that
I'm
leveraging
and
then
on
that
connection,
which
can
be
established
many
different
ways.
It
could
actually
be
a
point
to
site.
So
it's
just
my
particular
machine
gets
that
connection.
It
could
be
a
site
to
site.
So
it's
using
an
internet
connection,
but
it
creates
this
encrypted
tunnel
between
locations.
A
We
can
actually
be
a
direct,
a
direct
might
be
I,
get
dedicated
cables
or
I.
Have
a
software
defined
Network
an
sd-wan,
an
mpls
circuit,
to
connect
things
together.
This
can
become
an
offset
of
that,
and
so
that's
actually
a
better
way
to
get
that
connectivity.
So
now
the
upshot
of
all
of
this
is
it's
a
private
communication
to
my
resources.
A
I
mentioned
the
idea
that
all
of
the
resources
were
pulled
together.
I,
don't
just
have
these
individual
racks
that
have
a
finite
amount
of
capacity
that
I
have
to
pre-book
and
plan
that
ahead.
If
you
ever
use
a
cloud
service,
it
says:
well,
we
need
two
months.
No,
yes,
it's
like
it's
not
cloud
service.
A
They
have
millions
and
millions
of
servers,
they
pull
it
all
together.
So
there's
generally
not
a
limit
to
how
much
resource
I
can
create.
There
might
be
quotas
there
to
help
protect
you
from
yourself,
but
I
can
generally
go
and
either
get
those
expanded,
or
maybe
there
is
some
construct
like
a
subscription
or
an
account
all
right.
A
But
why
that's
so
useful
is
this
elasticity
means
I
can
do
things
like
scale
and
auto
scale,
it's
even
nicer.
So
what
that
means
is
I
can
horizontally
I.E
at
and
remove
instances
of
my
service,
because,
if
you
think
about
maybe
I
need
10
instances
of
a
web
service.
Well,
if
it's
a
quiet
time,
maybe
you
only
need
three
instances:
I
don't
want
10
running
because
we're
going
to
see
I
pay
for
what
I
use
so
as
the
load
changes
coming
into
this.
A
If
the
load
goes
up,
hey
add
more
instances
I'm
provisioning
on
demand
when
I
need
it.
If
the
load
goes
down,
I
remove
it
that
might
be
a
planned
scaling.
Hey
I
know
I'm
busy
at
8am
till
5
p.m.
I
am
selling
Pizza
on
Super,
Bowl
I
know
I'm
going
to
be
really
really
busy
or
I'm
selling
Pizza
on
it's
Friday
night
I
know
I'm
gonna
be
really
busy.
It's
tax
season,
that's
an
annual
seasonality.
I
know
I
need
more
resource,
then
maybe
it's
not
predictable,
but
I'll
always
be
able
to
react
to
that.
A
A
Use
I
pay
for
what
I
use
if
it's
a
VM
I
pay
for
the
seconds
that
VM
is
turned
on.
If
it's
storage,
maybe
hey
it's
the
amount
of
capacity
I'm
using
or
maybe
it's
the
amount
of
performance,
I've
provisioned.
If
it's
a
database,
it
might
be
a
function
of
hey
the
amount
of
space
I'm
using
and
the
amount
of
transactions
I
run
against
it.
If
it's
a
GPT,
how
many
tokens
are
being
input
and
then
output?
A
It's
always
going
to
be
a
function
of
what
I'm
using.
So,
if
I
think
about
this
provision
on
demand
and
this
elasticity,
this
measured
use
and
consumption-based
billing
I
only
pay
for
what
I
use.
That's
why
this
scale
idea
is
so
important
at
any
moment.
I
only
pay
for
what
I
need
at
that
moment
in
time
and
I
can
change
it
as
I
need
to
on
premise:
I
can't
do
that
on
premises.
I
I
have
to
buy
the
servers
and
remember
when
I
buy
a
server
I
have
to
think
about.
A
Well,
how
big
might
this
service
get
I'm
going
to
use
this
for
five
years?
I
should
buy
a
server
big
enough
to
support
the
next
five
years
so
end
up
with
a
huge
upfront
capital
expenditure,
as
you
may
over
buy,
which
is
why
VMS
were
kind
of
popular
and
help
solve
some
of
that,
but
they
still
have
to
buy
the
capacity
up
front.
I
have
to
buy
a
storage
area
network
Big
Lots
of
disks
up
front
I,
buy.
A
Front
I
need
technical
expertise
and
architecture
and
planning,
but
in
terms
of
expenditure
of
money,
if
I'm
a
startup
I
can
start
really
small.
As
my
business
grows,
I
grow
the
amount
of
resource,
it's
fantastic
for
me,
if
I'm
starting
a
business.
Well,
it
doesn't
work
out.
It's
that
idea
of
fail
fast
well.
At
least
I
didn't
waste
a
whole
bunch
of
money
buying
equipment
that
I
never
needed.
If
I
have
a
short-term
project
house
spin,
stuff
up
and
I
can
spin
up
ridiculous
amounts
of
power.
A
A
A
A
A
There's
going
to
be
governance,
there's
going
to
be
policy
that
can
control
what
they
can
do
where
they
can
create
it.
The
types
of
resources
so
I
don't
have
someone
create
in
a
storage
resource
and
exposing
it
to
the
public
internet
and
then
exposing
my
company?
No,
we
use
policy
to
enforce
the
requirements
we
have
in
our
company.
So
if
you
think
historically,
the
order
was
hey,
I
need
something,
I
go
and
ask
the
IT
team.
A
They
check
my
request
to
see
if
it
meets
requirements
and
then
they'll
action
it
well
now
the
IT
team
put
in
the
governance
controls
first
and
then
they
just
automatically
enforced
as
the
users
now
try
and
create
their
resources,
so
I
still
have
all
of
the
controls.
I
can
control
who
can
create?
Who
can
maybe
stop
and
start
who
can
scale
things,
but
is
the
IT
department
no
longer
have
to
be
involved
in
all
of
those
different
actions?
A
It's
very
granular
governance,
controls
and
policy
that
help
enforce
the
various
things,
and
these
are
really
the
key
aspects.
When
I
think
about
the
cloud.
This
idea
of
multi-tenancy
the
idea
that
it's
publicly
this
broad
network
access
I
can
provision
on
demand
of
amazing
elasticity
is
a
measured
service.
I
only
pay
for
what
I
use
and
hey
controlled
self-service
is
still
within
the
quotas.
Within
the
policies.
The
governance
controls
I
have
set.
A
And
that
I
mean
that's
the
cloud:
that's
the
public
Cloud,
it's
not
some
magical
thing!
It's
the
idea
that
now
any
company
has
the
ability
to
Leverage
ridiculous
amounts
of
capacity
through
a
massive
range
of
resources
that
enable
me
to
only
be
responsible
for
the
things.
I
want
to
be
responsible
for
I.
A
The
more
we
think
about
it
today,
the
more
we
think
about
hey
people
trying
to
do
harm
against
us.
These
companies
spend
huge
amounts
of
money
to
help
secure
this
and
again
I'm
more
focused
on
the
Microsoft
stack.
So,
yes,
we
have
these
fantastic
capabilities
in
the
cloud,
but
hybrid
is
a
big
thing:
hey,
maybe
I
need
to
be
able
to
use
these
services,
and
a
region
doesn't
work
for
me.
A
Maybe
there's
an
anchor
some
Mainframe
that
I
have
to
be
able
to
talk
to
or
some
regulatory
requirement
that
I
needed
in
this
location
or
there
isn't
a
region
where
I
need
it.
I
can
bring
these
services
to
my
own
stuff
to
my
own
service
to
my
own
locations,
so
I
still
maintain
that
flexibility,
so
I
hope
that
was
useful.
A
very
high
level
of
what
is
the
public
cloud.