►
From YouTube: OpenStack Austin Meetup 09 Jan 13
Description
Everett Toews (@everett_toews) give sponsor intro by Rackspace
Miguel Lavalle presents about Quantum (starts about 11 mins in)
Slides!: http://www.slideshare.net/MiguelLavalle/network-virtualization-with-open-stack-quantum.
REQUEST TO WATCHERS...PLEASE ADD COMMENTS WITH SLIDE X = MM:SS!
A
A
So
I'll
do
a
quick
introduction
that
will
turn
it
over
a
sponsor
track.
Space
welcome
to
I
robbers
felt
welcome
to
the
OpenStack
Austin
meet
up.
We
have.
This
is
a
monthly
event
where
OpenStack
people
get
together.
It's
dell
has
been
you
can
come
into
this
and
we
help
out
the
space,
and
then
we
share
sponsorship
with
different
companies,
and
we
have
a
bit
of
a
list
waiting
list.
A
If
you're
interested
in
sponsoring,
please
contact
me
robbers
felt
or
just
go
through
the
meetup
group,
jean-marie
martini,
who
helps
manage
the
sponsorships,
will
help
you
find
a
spot
sponsored,
and
so
we
continually
have.
If
you
want
talk,
if
you
have
topics
of
interest,
if
you
want
tavis
the
speech
or
if
you
know
of
an
openstack
event,
the
area
that
people
are
all
going
through,
please
go
through
to
make
it
up
side
view
and
we'll
work
with
you
to
do
that.
A
Tonight's
agenda.
Is
it's
going
to
be
a
short
presentation
with
Pratt
fire
response
at
Rackspace
and
then
be
else
going
to
give
a
introduction
to
quantum,
so
that
should
be
really
exciting
and
then
because
the
election
is
coming
up
and
I
know,
there
are
at
least
two
people
who
have
board
or
nominated
for
the
board.
We
will
at
least
stay.
A
Let
us
be
recognized
so
we'll
do
that
I
guess
it's
been
grabbed
up
again.
Take
next
agenda
items.
The
next
sponsor
I
usually
like
to
give
them,
since
early
props
is
going
to
be
SolidFire,
so
give
it
in
the
active
in
the
silver
any
questions
before
I
turn
it
up
we're
going
to
Wi-Fi
password.
There
should
be
an
open
one
that
doesn't
require
an
answer,
because
it's
a
great
kids,
take
it
different.
Yes,
okay,
oh
and
the
next,
the
next
meetup
brandon
t
know
in
the
next
ones.
It's
going
to
be
able
to.
C
B
Hi
everyone,
my
name,
is
Everett
tapes,
I'm,
a
developer
advocate
in
the
developer
relations
group
at
Rackspace.
So
it's
a
bit
of
a
mouthful.
What
is
being
really
all
developer
advocates.
Do
our
mission
is
the
mill
to
develop
a
relationship
is
to
make
it
easier
for
developers
to
use
the
cloud?
That's
it
period
and
probably
of
course
says
most
of
you
know,
making
something
easy
to
use
is
actually
really
really
difficult.
So
it's
a
lot.
B
First
of
all
can
I
get
maybe
by
a
show
of
hands
how
many
people,
here
our
developers
write
code
about
a
quarter
so
maybe
or
more
of
a
percent.
Great
ok.
So
the
developer
part
of
developer
advocate
is
I'm
a
software
developer,
but
I
also
represent
external
software
developers
to
backspace
within
Rackspace.
So
if
there
are
concerns
that
developers
have,
who
are
you
know
external
to
rackspace
outside
of
rackspace,
but
developing
on
the
Rackspace
cloud?
I
want
to
hear
those
concerns.
B
I
want
to
find
out
what
would
make
it
easier
for
you
guys
to
develop
your
applications
on
top
of
the
Rackspace
cloud,
it
kinds
of
things
you're.
Looking
for
many
grievances,
you
have
feedback,
positive
or
negative,
is
always
great.
That's
what
we,
what
we
feed
on
so
right
now,
one
of
the
first
things
we're
doing
to
help
developers
is
to
create
some
what
someone
has
SDKs
or
software
development
kit.
B
So
the
only
thing
I'd
ask
you
to
remember
from
this
most
feel:
is
dogs
Rackspace
calm
and
then
you
can
scroll
to
the
bottom
and
find
our
little
box
right
there,
the
SDKs,
the
software
development
kits
now
what
our
SDKs
they're
actually
comprised
of
a
few
things
there's
a
language
binding.
So
it
allows
you
to
program
against
the
Rackspace
cloud
in
the
language
of
your
choice.
B
So
if
you're,
a
Java
developer
or
a
Python
developer,
PHP
developer
and
more
to
come,
you
know
you'll
be
able
to
use
those
languages,
leverage
the
skills
that
you
already
know
to
develop
your
applications
on
top
of
the
Rackspace
cloud.
So
essentially
these
SDKs
here
make
it
easier
for
you
to
develop
your
applications
on
top
of
the
Rackspace
cloud
and.
B
Three
Java
PHP
and
Python,
we're
looking
to
add
more
ruby,
is
in
the
works
right
now
and
we're
looking
to
add
mobile
SDKs
as
well
for
mobile
applications
on
Android
and
iOS.
So,
for
example,
let's
take
a
look
at
the
Java
SDK,
which
is
the
one
that
I'm
currently
working
on.
It's
actually
based
on
an
open
source
project
known
as
Jay
clouds,
and
this
is
a
cross
cloud
toolkit
that
actually
works
across
a
number
of
clouds,
including
amazon,
HP,
google
compute
engine,
his
support
for
that
is
just
about
in
there.
B
So
you
know
why
would
we
choose
an
sdk
that
works
across
all
these
different
clouds?
What's
the
advantage
to
rackspace
for
doing
that,
you
know:
what's
the
benefit
to
you,
the
developers
for
doing
that?
Well,
the
point
is
you
know
we
want
you
to
be
able
to
port
your
applications
around
on
the
clouds.
B
So
there's
a
handful
of
things
involved
in
each
SDK.
Of
course
you
know
the
SDK
itself,
the
language
finding
that
he
can
download
and
install
the
kind
of
documentation
that
goes
along
with
that.
So
it's
it's
not
just
enough
to
throw
some
cold
over
the
wall
and
say:
oh
here,
it
is
use
this.
You
have
to
tell
people
how
to
use
it.
B
So,
of
course,
we
have
no
getting
started
documentation,
detailed
API
documentation
for
when
you
want
to
get
down
into
the
details
and
start
doing
very
specific
things
with
a
different,
different
api's
and
lastly,
some
sample
code
example
code
that
can
get.
You
started
right
away
that
you
can
use
as
kind
of
a
template
to
see
how
things
are
done
and
kind
of
a
best
practices
matter,
and
that
is
pretty
much
it
not
really.
A
presentation.
B
B
Absolutely
yeah
so
as
part
of
that
cross
cloud
toolkit
aspect
and
that's
this
one's
true
for
you
know:
Python
PHP,
J,
pods
Ruby,
coming
up!
Thank
you
Joe,
a
good
point.
They
all
do
work
across
OpenStack,
just
as
well
as
if
you
Rackspace,
and
you
can
write
the
same
code
and
you'll
run
both
on
rackspace
and
on
OpenStack.
Go.
B
Cipal,
of
course,
is
different
clouds,
whether
they're,
private
or
public
you're
going
to
have
different
capabilities.
So
you
actually
can
the
typically
the
EP
eyes
are
covered
right
down
to
the
very
ground.
But
if
there
isn't
a,
if
it's
not
common
functionality,
then
it
will
be
part
of
that
portable
layer
that
you
would
write
your
code
to
you
actually
have
to
say:
okay
well,
I,
take
a
step
down.
I
start
writing
to
specific
OpenStack
code.
B
F
B
So
in
particular,
Jay
clouds
I
can
speak
to
specifically.
I
know
that
when
it's
so,
for
instance,
like
the
security
groups,
api
is
considered
an
extension.
So
within
j
clouds
you
can
actually
query
API
itself
the
j
cause
idea
and
say
you
know,
does
this
extension
exist?
If
so
use
it,
you
know,
and
then
you
can
write
your
code
to
manipulate
security
groups.
If
not,
you
know
maybe
use
another
mechanism
just
discover
exactly
yeah.
H
H
All
right,
my
name
is
a
again
Hulk
and
we're
going
to
talk
about
the
never
be
Tony
station
series
of
OpenStack
I've
been
a
quantum
developer
since
about
DM
October
right
now,
I
live
in
the
effort
to
develop
the
functional
tests
for
that
for
the
point
and
servicing
in
OpenStack.
If
you
have
traveled
with
my
lips
out
exercising,
please
I
can
see
they
to
raise
your
hand
and
I'll
be
happy
to
repeat
wherever,
whatever
we
haven't
really.
H
This
is
the
elder
law
outline
of
current
presentation.
We
are
going
to
talk
about
a
little
bit
of
our
quantum
suites
in
the
overall
OpenStack
iron
landscape,
I'm,
making
the
assumption
that
some
of
you
are
not
really
you
don't
you
don't
really
know
much
about
open
structure,
I'm
going
to
assume
that
we
need
to
go
from
the
meaning
and
that's
why
I'm,
including
these
five,
the
presentation?
H
Then
we
are
going
to
talk
about
white
white
quantum.
What
why
are
the
two
fundamental
reasons
of
having
quantum
OpenStack?
We
are
going
to
talk
about
an
overview
of
the
API.
What
are
the
main
up?
Abstractions
of
the
API
offered
to
the
users
of
the
plan.
We
are
going
to
talk
about
the
plug-in
architecture,
how
that
API
is
implemented
through
a
plug-in
architecture,
and
then
we
are
going
to
show
you
a
demo.
H
We
are
then
going
to
talk
about
how
those
virtual
networks
are
realized.
Physically
in
the
environment,
we
are
going
to
talk
about
one
specific,
a
plugin
which
is
a
open,
the
switch
and
then
I'm
going
to
share
with
you
their
way.
I
set
up
this
the
environment
for
this
cloud.
My
questions
give
you
so
far.
H
H
We
have
a
young,
we
have
a
a
job
application,
which
is
a
console
risin,
presently
a
web-based
interface
for
OpenStack,
and
there
are
several
clients,
CLI
clients
for
the
different
services.
So
there's
there's
one
cl
a
client
for
bible
and
they're
sincere,
like
lying
for
the
other
senses,
we
have
a
dealer
serving
the
other.
More
important
services
are
glands,
which
is
a
eat.
H
My
services
does
for
users
and
tenants
can
upload
their
images
and
when
you
are
creating
virtual
machine
in
ANOVA,
you
are
really
creating
an
instance
that
virtual
machine
with
one
of
the
images
implants.
Okay,
we
have
the
sweet
service,
which
is
an
object,
store
service
similar
to
Amazon,
cestus
3,
and
that's
where
you
can
store
long-term
storage
for
objects.
H
In
the
plan
we
have
key
stone,
which
is
the
authentication
service,
in
other
words,
when
a
tenant
wants
wants
to
use
access
the
openness
type
while
they
are
they,
inter
with
Keystone
and
that's
how
they
authenticate
their
arrested,
something
that
I'm
not
showing
here
because
I
didn't
have
any
more
spaces,
have
not
very
important
service
cinder
and
cinder.
Is
they
they
block
stores
services
that
that
service
was?
It
was
originally
part
of
Nova
and
it
was.
It
was
separate.
It
was
putting
a
separate
service
since
the
current
version
of
open,
strap,
which
is
awesome.
H
H
Again
for
those
of
you
of
you
who
are
not
familiar
with
the
OpenStack,
we
are
right
now
we
are
rough
that
the
current
releases
fall.
So
we
decided
our
releases
by
letters
and
the
previous
release
was
a
sex
clear.
We
have
the
current
stable
releases
folsom
and
we
are
working
on
the
result.
I've
risen
in.
H
H
Separating
the
service
we'll
get
to
that
in
a
Pollock
sighs
and
if
we
thought
that
will
get
back
to
you
to
your
questions,
so
so
why
quantum
number
one
we
want
to
give
up
tenants
who
want
to
give
uses
full
control
of
their
network
environment.
In
other
words,
we
want
if
we,
if
we
are
making
the
assumption
that
that
customers
and
tenants
are
going
to
move
enterprise
applications
to
open
sky
clouds
will
want
them
to
be.
We
want.
H
We
want
them
to
be
able
to
replicate
whatever
network
architecture
they're
coming
plates,
regardless
of
how
those
networks
are
realized
physically
in
an
open
structure.
So
we
wanted
to
be
a
full
give
them
full
freedom.
And
that's
that's
why
that's
what
we
mean
by
by
tenant
control
one
day
we
want
to
give
them
full
control
of
their
networking
bar
in
their
applications.
We
want.
We
want
the
application
to
define
what
the
networking
architecture
is.
We
don't
want
the
cloud
to
the
digging.
H
So
right
now,
then,
tendons
are
able
to
create
multiple,
multiple
private
l2
networks.
They
are,
they
think
they
have
full
control
of
their
IP,
addressing
your
needs.
They
cannot
monitor,
basically
nebulous
network
status
and
they
can
connect
to
router
services
and
create
on
double
those
are
into
networks,
more
complex
services
to
build
three
or
four
level.
H
The
other,
the
other
reason,
the
other
justification
for
a
quantum,
is
to
give
determines
that
the
freedom
to
use
whatever
networking
technology
is
best
for
further
application.
We
want.
We
want
the
virtual
networks
to
be
completing
the
payment
of
how
those
networks
are
implemented
from
a
physical
physical
technology
point
of
view.
So
the
way
these
circuits
and
at
this
before
before
quantum
Noah
had
was
a
limited
to
basic
villain
villain
networks
and
Linux
I
iptables.
That's
all
there
was
for
tenancy.
H
So,
if,
if
they
want
to
realize
their
their
work
net
worth
using
tunneling
sowwy
SDM
VPN
solutions
help
to
fabric
solutions,
and
whatever
other
words,
would
you
you
want
to
incorporate
the
two
to
the
club?
The
idea
is
that
that
we
have
a
plug-in
architecture
and
through
that
plug-in
architecture,
the
users
and
the
talents
are
able
to
connect
whatever
physical,
never
gain.
Technology
suits
their
opinions.
H
H
H
C
C
C
H
So,
what's
what's,
what's
it
the
base?
He
gave
you
up
api
abstraction
if
we
make
a
parallel,
the
matter
with
noah,
with
the
computer
visit,
the
main
obstructionism
virtual
mention
that
it's
connected
to
something.
So
that's
that's
the
main
attraction
law
provided
by
the
compute
service.
For
the
total
mountain
for
the
networking
services,
we
have
three
main
instructions,
number
one
we
have
as
I
said,
L
to
level
two
virtual
network.
That's
that's
one
key
abstraction.
H
I'm.
Sorry,
if
you
read
their
the
quantum
API
manual,
you'll,
see
that
all
that
is
offered
by
the
API
is
operations
to
manipulate
those
three
abstractions.
You
create
networks,
you
live
networks,
you
create
sub
networks
and
create
a
virtual
box,
so
it
should
suggest,
implement
it
right.
The
early
we
will
get
them.
Yeah
bet
you
got
an
essentially
and
essentially
the
idea
is
that
using
these
basic
components,
you
can
assemble
more
complex
services
of
more
complex
tools
to
create.
H
So,
using
these
getting
back
to
your
question
using
these
basic
services,
then
you
create
more
complex
or
environments
on
that
landscape.
For
example,
you
can
you
say
you
can
create
hours,
you
can
create
external
networks,
you
can
create
their
floating
IPS
to
allow
access
to
it
to
the
vm
so
from
the
outside
and
so
on.
Ok,
and
the
fundamental
idea,
is
that
you
create
your
virtual
networks.
You
create
yourself
networks
and
you
connect
that
you're
settled
in
whatever
we
community
limitations
on
meaning
is
between.
E
H
H
So
from
frontlets,
a
very
important
also
be
used
to
make
a
distinction
between
the
10
and
bu
and
the
physical
p.
So
from
from
the
tenon
point
of
view,
you
have
that
you
have
the
virtual
net
words,
you
have
your
own
addressing
spaces
and
they
are.
They
are
completely
isolated
from
other
tenants
and
then
the
way
those
physical
networks,
I
realized,
realized
in
the
in
the
in
the
cloud.
H
H
Another
important
concept
here
is
what
is
the
API
extension
so
there's
a
basic
API
person
named
API,
which
offers
the
three
abstraction
that
we
already
talked
about,
but
also
their
said
their
say,
you're
a
very
deliberate
deliberate
effort
to
create
API
extensions
I
get
the
idea
is
that
the
assumption
is
that
3d
extension
innovation
is
going
to
be
incorporated
to
not
only
that
some
of
those
have
extensions
that
become
widely
used
by
our
users.
Then
they
are
incorporated
eventually
over
time
within
the
main
NP.
Ok,.
H
So,
just
as
a
summary,
the
current
currently
these
are
the
abstractions.
There
are
available
in
the
folsom
release,
so
you
have
networks
you
can
submit
you
have
towards
and
right
now
we
have
routers
and
floating
eye
peas,
SMS
an
API
extension
and
actually
get
the
plan
for
grizzly
that
these
abstractions
are
moving
to
the
to
the
main
ATM.
H
Networks
are
classified
in
in
two
dimensions:
internal
and
external,
so
starting
a
meaning
that
they
have
access
to
outside
the
cloud
and
they
can
be
accessed
from
outside
the
club,
an
internal
they
are
only
visual
within
within
the
do
construction,
and
you
have
private
networks,
and
you
have
sure
there
was
meaning
that
as
a
tenant,
I
can
I
can
create
a
real
word.
That
can
be
final
networking
and
say
that
this
is
this
is
solomon
is
only
for
my
application.
H
B
E
H
E
H
H
So,
what's
the
status
are
so
far
at
the
falsehood
release,
we
have
a
bit
version:
2
API,
with
those
abstractions
that
we
already
covered,
which
is
l2
close
iprs
management.
We
have
a
integration
with
Keystone.
Keystone
again
is
the
authentication
service
and
integration
with
horizon
with
the
GUI
interface.
We
have
a
command-line
interface.
I
wish
have
two
extensions,
which
are
every
flowers
with
floating
IPS.
H
H
H
H
And
these
are
the
commands
that
I
used
to
create
that
work.
A
cuanto
plan
to
net
create,
I
created
the
network
private,
and
I
did
that
barossa
as
an
admin,
so
I
had
to
specify
the
ID
for
the
tenant.
Okay,
I
created
a
subnet
with
this
IP
address
range.
I
created
the
world
and
then
I
connected
that
router.
To
that.
C
E
H
So
the
other,
the
other
piece
that
I
created
in
this
environment
is
an
external
network.
Okay,
with
this
IP
addressing
space,
and
the
idea
here
is
that
when
you
create
an
external
network,
these
IP
addresses
are
floating
and
you
assign
them.
You
can
assign
them
to
you
to
use
the
MS
in
the
end,
and
by
doing
that,
then
you
are
able
to
access
UVM's
from
from
out
slid
open
stuff
from
from
the
outer
world.
If
you
look,
does
that
make
sense?
H
That's
that's
a
reason
to
create
a
a
external
enemy,
the
other
from
from
the
point
of
view
of
the
of
the
dms
that
you
have
in
your
OpenStack,
the
router
and
the
external,
allow
those
VMS
to
access
that
the
absolute.
But
for
that
you
don't
need
you
don't
need
a
floating.
A
thing
you
need
are
floating
I'd,
be
only
for
the
outside
to
be
able
to
rush
the
Europeans
make
sense.
H
So
again
we
saw
the
commander
that
I
used
to
create
that
their
environment,
a
and
again
I
created
a
network
but
I
specify
that
that's
an
external
okay
and
really
for
me
from
an
implementation
point
of
view
and
external
network
is
a
network
that
it's
attached
to
a
bridge
called
be
be
our
X.
So
that's
external
breach.
It's
a
bridge
connected
to
an
interface
and
when
I
create
an
external
network
and
connect
to
my
network.
To
that
which
and
I'm
going
to
show
you,
then
I
create
a
subnet.
H
H
H
H
H
H
H
It's
as
simple
as
that
now
for
those
of
you
who
haven't
played
with
the
OpenStack
at
all
I'm,
going
to
query
glance
glance.
It
again
is
the
images
services,
image
images,
manager,
management,
services,
service,
I'm,
going
to
query
glance
as
to
what
images
are
available
in
this,
in
this
case
to
glance
a
match
list.
H
H
H
H
H
H
B
H
H
H
E
D
H
Ok,
so
now
we
are
going
to
a
boot
up
our
we're
going
to
greater
or
instances
in
this
case
and
instead
of
using
the
quantum
command
I'm
going
to
use
a
moment,
because
that's
a
that's
it
people
get
service.
So
it's
not
about
the
image.
Is
that
that
image
that
I
show
on
the
screen
a
few
minutes
ago,
the
flavor
flavor
one
and
the
flavor
is
that
the
size
of
my
BM
I
can
have
a
tiny
Veatch
is
for
one
I
have
a
small
which
is
flavor
two
three
and
so
ok.
B
H
Its
own,
it's
gonna,
if
I
make
a
mistake
like
that
turn
it
it's
gonna
know
is
going
to
come
back
and
say
no
I,
don't
find
that
resource
on
you're,
missing,
you're
missing
an
arm.
Ok,
ok!
So
now,
I
have
a
have
a
my
first
vm
and
again
it's
a
tiny
bien.
Ok,
so
very
little
small
DMD
useless
little
resources
as
possible
to
produce
them.
H
C
E
B
H
H
H
H
H
C
H
H
H
E
D
H
Interferes
ero
is
connected
to
10
points.
It
1000
watts
for
now
remember
that
this
vm
should
be
connected
to
two
networks,
the
reason
and
it's
connected
to
two
networks
that
the
reason
we
don't
see
it,
because
that
second
interface
is
not
configuring,
the
operating
system,
but
we
are
going
to
do
it
now.
So
I'm
going
to
say
solar
if
config
interface,
one.
H
H
H
H
C
H
H
H
You
can
see
that
that
well,
you
cannot
see
in
the
Bible.
Essentially
I
have
two
ports,
one
power
side,
091
was
trying
to
network,
and
I'm
going
to
use
the
ID
of
that
pork,
because
I'm
going
to
assign
a
floating
I
to
develop
or
so
I
can
access.
This
vm
from
the
outside
I
can
move
in.
Does
that
make
sense.
H
F
C
H
So
I
created
a
floating
IP
address
and
it's
it's
telling
me
that
that
my
floating
IP
address
is
172
though
3103
okay
and
that's
that's
in
the
range
of
my
offloading
IP
network
that
I
created
all
right
now
it
doesn't
have
a
fixed
IP
IP
address
because
I
created
the
floating
IP
address,
but
it's
not
a
something
to
any
any
vm.
So
next
step
is
I'm,
going
to
assign
Giza
floating
IP
to
the
port
that
I
displayed
in
my
previous
command.
Does
that
make
sense
because
that
poor
be
launched
to
my
vm?
H
H
H
Like
the
floating
idea
stays
with
the
end
for
as
long
as
I
listen
floating
IEP
side,
chlorine,
eyed
peas
are
scarce
resources
that
the
social
every
every
vm
gets
gets
internal
IP
addresses,
but
when
you
connect
them
to
two
levels,
but
flooring
IP
address
is
very
scarce
in
the
sense
that
they
have.
They
have
to
be
recognized
in
the
outside
world.
So
so
so
you
should
you
that's
why
you
assign
them
as
floating
IP
addresses.
Does
that
make
sense
yeah?
But
when
you
have,
when
you're
failing
a
vm.
A
H
A
H
A
H
G
E
G
G
D
H
H
Me
say
you
probably
probably
Maggie.
That's
part
of
the
problem
at
that
point
on
was
that
you
only
had
a
limited
range
of
options
to
realize
these
networks
from
a
physical
point
of
view,
whereas
in
quantum
there
is
a
whole
new
world
that
allows
you
to
incorporate
as
many
physical
networks
as
your
your
application
reports.
Now
again,
that's
going
to
depend
on
how
how
efficiently
you
Sarah,
but
but
your
local
train
anymore
by
the
by
the
fact
that
you
only
have
one
technology
option,
which
is
what
which
were
villains.
A
H
A
H
So
I'm
going
to
do
five
more
minutes
and
essentially
what
I'm
going
to
talk
is
how
you,
how
we
physically
realize
these
logical
network,
since
we
know,
but
I
can
essentially
what
you
do
is
you
have
a
garage,
achill
networks
and
those
musical
their
words
are
mapped
by
the
environment
to
to
physical
networks.
There's
a
there's,
a
a
API
extension
called
the
provider
extinction
and
that's
the
prob
I
extension
is.
He
said
the
portion
of
the
API
of
the
extended
API
that
allows
admins
to
mark
those
neighbors
to
the
physical
network.
Successful.
So
that's!
H
That's
that
the
provider
extension
and
with
this
extension
uses
with
admin
privileges
can
create
logical
networks
that
are
explicitly
mapped
to
physical
movements
and
there's
their
say.
Some
key
terms
to
understand
number
one
virtual
network
see
which
one
level
he
said.
We
shall
be
final,
two
segments
say
it's
an
l2
segment.
The
way
we
will
talk
about
them
so
far,
you
have
physical
networks,
which
are
networks
physically,
connecting
the
different
notes
in
the
in
the
open
starter.
In
an
instance,
you
have
10
and
met
with
a
tenant
network.
H
H
Now,
how
do
you,
how
do
you
specify
that
this
is
an
open
base
which
implementation
in
your
open
be
switch?
Configuration
file?
You
specify
that
the
tenant
networks
type
is
villain.
This
is
a
deal,
an
exam.
Okay,
you're
mapping,
your
your
net,
your
tenant
levels
to
villain,
you
configure
you,
your
open,
be
switched
agent
to
with
with
this
one,
and
also
you
specify
the
rangers
ranges
for
your
villain
and
essentially
what
you
do
is
you
say:
okay,
I
have
a
physical
network
that
has
this
virtual
ID
bring
this
range
available.
H
H
Of
course,
I
I
create
a
bridge,
a
physical
bridge
that
maps
to
an
interface
connector
in
each
Noah
had
to
create
a
bridge
that
connects
to
an
interface
connected
to
those
physical
levels,
ok
and
finally,
in
the
different
compute
nodes.
What
I
say
is
the
wave
those
physical
networks
map
two
interfaces,
so
please
in
compute
node
a
a
computer
is
a
not
running
VMS.
I
have
physical
network
one
map
to
interface,
one
and
physical
network
to
but
to
inter,
inter
phase
two
does
that
make
sense.
H
Correct
that
and
the
enemy
here,
I'm
Gavin,
but
the
key
point
here
is
that
I
specify
that
tended
networks
type
is
Dylan,
meaning
that
when
my
tenants
create
a
virtual
network
open
a
stacked
us
this
fur
for
them
or
not,
they'll
have
to
worry
about.
The
OpenStack
is
mapping
those
tenant
levels
to
those
physical
memories.
H
Id
ranges
the
range
of
tunnel
ideas,
what
I
want
to
use
and
I
in
each
compute
now
I
specified
the
end
point
where
the
tunnels
are
connected,
which
is
just
an
IP
address.
Ok
from
that
point
on,
my
virtual
networks
are
going
to
be
mapped
by
again
my
OpenStack,
my
tenant
networks
are
going
to
be
mapped
to
Jerry
terms.
H
H
F
C
H
B
Question
is
kind
of
for
the
whole
meet
up
here
in
pretty
general.
So
now
that
we've
got
software-defined
networking,
that
means
we
have
developers
creating
not
only
BMS
but
networks
as
well.
So
you
know
creating
bmz
developers
have
had
to
get
a
little
more
into
ops,
which
is
definitely
a
good
thing,
but
it
also
makes
it
more
dangerous
and
now
every.
B
Makes
it
even
more
dangerous,
like
I've,
got
a
relatively
thin
knowledge
of
networking
and
you
know
creating
networks,
and
you
know
specifically
with
security
implications.
You
know
I
just
I,
don't
know
all
the
bad
that
I
can
be
doing.
I,
don't
even
know.
You
know
what
the
dangers
are
of
if
I'm
configuring
a
network
in
one
way
versus
another.
So
like
what
the
skill
set
that
I
mean
developers
have
to
become.
You
know,
network
operators
now
all
right.
C
I
E
I
Like
some
of
the
stuff
isn't
supported,
maybe
because
they're
doing
IP,
addressing
and
they're
masking
I
can
use
to
go
from
in
trouble.
Exhale
I
would
hope
it
did.
Somebody
be
interesting
have
a
really
deep
that
would
be
interested
in
a
deeper
dive
on
exactly
how
those
networks
communicate
together,
because
it
God
has
surprised
security,
builder,
yeah,.
A
I
mean
you
could
do
an
equally
long
and
detailed
presentation
on
the
implementation,
which
is
most
appreciation
to
run
amok,
see
if
they
much
money
implementation
below
the
API
rather
than
this,
which
I
that's
like
the
implementation.
Above
me
it.
Yes,
one
of
the
trends
that
we've
seen
consistently
throughout
as
in
sex
life
cycle
and
I.
Expect
in
you
is
you
keep
cutting
pieces
out
of
open,
sat
and
and
basically
making
them
into
their
own
projects.
So
the
sum
of
these
shins
are
really
around
this
essence
of
OpenStack
Nova
started
with
the
networking
layer.
A
A
But
you
can't
just
off
immediately
switch
that's
hard
yeah,
that
was,
that
was
tricky
for
networking
for
storage.
We
actually
may
be
other
decision,
which
was
ok,
we're
going
to
move
the
codes.
It
does
storage
out
into
cinder
to
the
project.
For
that
and
then
make
that
more
complex,
we
did
not.
We
did
networking
a
little
bit
more
complex,
longer
slower
transition.
So
what
what
you're
asking
about
to
me
is
the
confusion
of
weight
notice
out
networking?
I
A
A
networking
is
deprecated
and
in
the
prisoners,
disease
you
know
what
we're
really
helping
people
do
is
start
migrating
towards
this
and
not
into
using
the
Nova
api's.
Mere
fact.
I
expect
providers
won't
even
offer
than
ever
it
networking
pieces
because
they're
much
less
configurable
for
the
end
user,
yeah.
C
D
At
I
want
I'm
from
security
point
of
view.
How
are
you
protecting
from
out
of
them
a
sneaker
other?
At
the
end,
the
beautiful
switch
is
thing
running
on
the
hypervisor
and
that
actually
activates
for
mr.
from
promiscuous
mode
by
the
ball.
Is
there
a
way
to
actually
protected
against
it
or
I
mean?
Because,
if
that's
the
case
doesn't
matter,
content
of
this
in
another
network,
you
can
just
put
out
therapy,
be
very
well
actually
process
everything.
So
the.
A
Hardest
part
of
what
this
is
designed
to
address
is
exactly
that
problem
in
most
clouds
you're
on
a
shared
network
with
all
the
other
tenants
right
unless
for
Nova,
unless
you
switch
into
VLAN
mode
today,
and
then
you
can
set
up
dedicate
s
and
run
out
of
he
lands
yeah
so
thought
commercial.
What
this
does
this
is
play.
A
H
C
H
D
I'm
talking
about
like
for
instance
ago,
widow
candies,
which
are
the
inspiration
different,
but
at
the
end,
is
pretty
much
a
semi
of
the
colonel
is
the
one
actually
doing.
The
switching
in
there
I
mean
will
activate
the
hypervisor
I
mean
on
the
host
of
disco.
Balls
will
activate
at
us
/
mistress
move
in
order
to
pass
over
traffic.
From
one
point
we
have
right.
D
Pretty
much
the
same
stuff
with
the
performance
hit
unless
you
actually
activate
at
least
on
this
server
besar
ioe
and
put
on
top
of
that
running
of
your
bio,
you
won't
get.
You
will
still
lose
to
thirty
percent
of
performances
that
you
get
with
any
relation,
so
I
mean.
Does
it
does
those
sort
of
questions
that
are
like
that
out
there
on
earth
it
and
that's
one
of
the
main
reasons
why
at
least
an
hour
you
know
Tony
guards
were
still
having
a
great
view
on
committing
either
visualization
club
or
anything
else
like
that.
D
A
A
H
So
to
find
out
things
number
one
if
you
for
the
group
feel
that
you
want
to
do
a
follow-up
session
really
more
of
an
implementation
point
of
view
of
beatitude
to
explore
that,
because
I'm
doing
any
way
we
can,
we
can
even
explore
one
of
the
easier
load
flow
plugins
instead
of
option
to
switch
girl.
So
that's
point
number
one
and
number
two
number
two
just
to
close
I
retire
from
HP
a
few
months
ago.
So
I'm
looking
for
a
job.
A
So
wow
it's
been
a
great
crowd.
Please
tweet
respond!
Do
whatever
post
your
comment
since
the
thing
I
hope
you
come
back,
this
quantum
is
a
great
great
topic.
We'll
have
more
on
them.
I
know
we
have
some
centerpieces
coming
up
we're
going
to
try
to
bring
it
also
sixers,
as
things
get
bigger
and
bigger,
they
get
more
exciting,
but
next
up
its
iconic
is
in
april,
and
so
our
normal
schedule
will
be
will
start
talking
about
the
summit
for
a
number
with
that.
A
Hopefully
everybody
here
is
a
member
of
the
OpenStack
foundation,
the
free
thing
to
do.
If
not,
I
can
post
a
link
for
you
to
join
if
you
have
been
a
member
of
the
foundation
for
more
than
six
months,
which
is
a
pretty
tricky
thing
to
do
so.
That
means,
if
you
join
the
foundation
before
the
first
election
and
you
are
eligible
to
vote
for
the
upcoming
election,
which
starts
next
week.
So
I
think-
and
you
didn't-
you
didn't
actually
fill
it
finish
nomination
or
you
know
right.
A
A
Get
you
get
eight
votes
and
then
you
can
pass
them.
Probably
won't.
We
have
an
excellent
slate
of
candidates,
so
you
know
I'm,
actually
I'm
fully
behind
every
single
person
on
that
list,
because
I
know
them
all
personally,
they're
all
lots
of
people,
but
please
vote
right.
You
know
both
both
the
people
for
the
people.
You
want
about
how
you
know,
however,
you
do
it,
but
it's
really
important
for
Austin
American
to
make
sure
that
our
voice
is
represented
in
technology
stack
is
this
turnout
is
great.
A
D
A
Works
over
there's
various
there's,
actually
a
really
good
luigi
page
on
open
site
governance.
There
are
technical
committees
and
user
communities
and
the
board,
and
we
have
an
amazing
community
process
and
the
foundation
itself
has
funded
foundation
a
four
by
membership,
corporate
memberships
and
that
has
staff
in
austin
they're,
actually
headquartered
in
austin
and
in
most
of
the
staff
lives
and
work
there.
So,
but
I
encourage
you
to
go
check
ticket.
It's
actually
great
document
issues
done
a
good
job,
doing
that,
but
we're
always
evolving
how
we
do
it.