►
From YouTube: An Open Source Telco Cloud Stack to Accelerate the Adoption of... Nevicato Guillaume & Mathieu Rohon
Description
An Open Source Telco Cloud Stack to Accelerate the Adoption of Cloud Native Operator Model - Nevicato Guillaume & Mathieu Rohon, Orange
The Telco Cloud stack project showcases the value of a cloud-native, scalable, and automated Cloud infra. This Telco Cloud stack is built with open-source components and it fulfills the telco requirements. First Part, present the goal of the project to release a telco cloud stack to support 5G CNF (Cloud-native Network Functions) : Open RAN & 5G Core CNF. How will this initiative contribute to foster Cloud native infra for Telco purposes ? How will this initiative be interlocked with an existing ecosystem such as LF ANUKET or the O-RAN Alliance ? Why did the project choose to be Open Source released ? In which existing communities the development team will also directly contribute ? Second Part : demo based on a distributed Cloud Native federated lab across different Operators with several CNFs. The demo starts with an ANUKET conformance test of this cloud native infra as it is used to guarantee the conformance of the infra with telco requirements. Then we demonstrate a CNF Test suite of the CNCF, as it is a key verification to guarantee an efficient LifeCycleManagement TBC : demo showcases of Bare Metal opensource including Metal3 & Ironic to the Rancher stack.
A
A
The
second
topic
is
that
there
were
some
crucial
security
incident
last
year
and
the
things
are
going
worse
and
worse
for
the
operator.
They
are
more
and
more
the
target
or
on
some
cyber
security,
and
here
I'm
just
referring
about
a
report
of
the
gsma
for
this
year
and
last
point
is
that
operators
are
quite
convinced
of
the
benefits
of
a
cloud
native
and
the
good
things
is
also
that
some
network
functions.
Sorry,
some
network
functions
are
now
also
integrating
the
cloud
native
design
and
required
some
efficient
cast
to
power.
A
This
network
function
the
last
things,
and
here
is
also
an
echo
of
mihal
presentation
from
dt
this
morning.
We
need
to
have
continuous
change,
maybe
small
change
yeah,
but
we
definitely
need
to
have
on
a
flow
of
changes
on
our
infrastructure
and
on
our
operator
and
it's
very
important
to
maintain
innovation,
but
also
for
security
perspective.
A
So
this
is
the
situation
today.
Our
proposal
is
to
build
a
taco
cloud
stack
and
we
want
also
to
to
to
to
publish
in
an
open
source
way.
We
identify
the
different
pillars:
the
first
pillars.
We
we
had
a
tremendous
talk
before
concerning
performance.
Yes,
performance
is
really
important
in
the
native
environment,
with
a
different
type
of
optimization
mechanism,
dpdk,
bpf
and
and
so
on.
We
also
need
low
latency
when
we
want
to
tackle
some
use
cases
like
open,
running
and
performance.
A
Also,
it
has
been
mentioned
before
to
have
the
best
performance
we
need
cast
on
bar
metal
to
avoid
a
hypervisor
layer,
something
new
also
in
the
telco
cloud
infrastructure.
Is
that
really
we
need
distributed
cloud
and
to
manage
this
distributed
cloud?
We
have
two
different
layers:
the
first
layer
is
the
bar
metal
automation
or
I
can
manage
thousands
of
nodes
with
open
operating
system,
lifecycle
management,
for
example,
and
but
also
to
onboard
the
declarative
and
github
approach
to
to
manage
such
infrastructure.
This
is
a
very
important
point.
A
I
mentioned
that
security
is
a
must-have,
and
here
we
want
really
to
to
consider
that
security
is
an
added
value
of
this
project.
It's
not
something
that
is
going
to
slow
down
us,
so
we'll
integrate
this
as
a
core
element
of
this
project
and
based
also
with
some
some
eu
some
eu
incomes
about
what
can
be
the
security
design
on
on
such
infrastructure.
A
We
mention
this,
but
when
you
want
to
have
a
multi-vendor
approach,
it's
really
an
important
topic.
So
this
is
our
pillars
for
projects.
So
now
we
want
to
build
this
thing,
so
how
we,
how
we
are
going
to
to
to
have
it.
So
really,
as
I
mentioned,
we
we
didn't
just.
We
did
not
just
want
to
have
a
another
stack.
We,
our
focus,
is
to
have
a
production
grade
release,
so
it's
important
to
to
to
work
with
the
cooperators
we
we
are.
A
We
are
many
of
the
cooperators
that
need
this.
This
this
assets
and
the
it's
important
also
to
onboard
some
vienna
vendor
to
integrate
their
requirements
as
soon
as
possible
in
the
product,
lifecycle
management.
So
as
partners
of
course
orange
that
are
here,
but
also
telecom,
italia,
dutch
telecom,
telefonica
vodafone
nokia
and
ericsson.
A
We,
we
are
commonly
working
on
this
project.
We
are
about
30
engineers
that
have
been
already
on
board
on
this
project
and
you
have
some
first
assets,
but
we
want
also
to
accelerate
to
recruit
more
from
these
members,
but
also
to
open
this
project
more
and
to
to
have
a
key
contributor
to
to
move
the
market.
Here
is
the
main
goal
is
that
we
we
think
that
alone
it
would
be
very
complex,
so
we
really
want
to
to
have
an
alliance
to
to
move
this
market.
A
A
Definitely
we
have
some
members
and
some
engineers
that
are
fully
involved
also
in
open
run
alliance
and
the
war
group
6
concerning
infrastructure
is
really
key
on
this
initiative
and
maybe
on
the
future.
We
we
we
should
work
more
closely
with
gaiax
when
we
will
have
some
some
work
and
some
some
output
from
the
infrastructure
stream
of
gaiax.
A
A
I
already
mentioned
the
bar
metal
automation.
That
is
really
key.
We
want
also
to
go
in
the
network
automation.
We
spoke
with
different
engineers
before
during
lunchtime
about
the
the
nightmare
of
managing
lots
of
vlan.
So
we
really
need
to
to
optimize
this
part
and
hardware
acceleration.
It's
a
key
topic.
B
B
B
We
are
seeing
that,
especially
for
the
run
use
case.
The
bare
metal
management
part
is
really
important,
because
for
the
moment
we
are
seeing
that
run,
vendors
are
coming
with
with
strong
requirements
about
the
the
os
and
the
and
the
materials
running
below
the
kubernetes
stack.
So
we
need
to
find
a
way
to
operate.
This
stack
as
we
do
operate
the
cnf
on
top
of
it.
B
B
A
Yeah-
and
it's
really
important-
that
we
are
based
on
open
source
components,
we
will
contribute
also
on
this
open
source
component
and
we
will
release.
We
will
release
and
publish
this
integration
in
an
open
source
way,
and
it's
really
important
to
to
do
this,
and
for
so
that
this
telco
cloud
stack
can
host
these
different
use.
Cases
that
you
can
see
on
top
open
on
is
an
example.
5D
core
is
an
example,
but
we
do
not
want
to
limit
a
boost
cnf.
We
want
also
to
address
the
cdn,
the
sd1.
A
B
This
demo
was
demonstrating
how
we
can
connect
labs
from
from
all
the
partners,
also
telco
partners,
dutch
telecom
telecom,
italia,
telefonica
and
orange.
On
those
labs.
We
were
deploying
a
management
stack
based
on
the
terraform
script
available
to
deploy
the
stack.
The
management
stack
were
deployed
on
the
in
the
orange
lab.
B
B
So
this
is
an
integration
project
where,
where
the
sea
highs
will
be
really
important
in
order
to
have
a
production
grade
stack
and
we
are
going
to
rely
on
existing
testing
frameworks
such
as
anuket,
rc2,
cnf,
conformance
test
suits,
sonoboy
and
so
on.
This
kind
of
test
suit
will
be
included
in
our
ci
and
we
will,
depending
on
on
how
it
works.
We
will
include
them
in
the
in
the
ci
or
not,
but
if
not,
we
will
have
a
deep
understanding
of
the
reason
why
it
is
not
working.
B
It
is
a
cloud
native
application
developed
by
nokia
and
we
work
together
with
a
telecommittee
earlier
devops
team
and
a
nokia
devops
team
to
to
be
able
to
deploy
it
with
with
fleet.
So
with
the
github's
approach,
be
able
to
scale
the
the
components
provided
by
the
the
cnf
scale
them
up
and
down,
and
we
will
show
this
in
a
in
a
in
a
video.
B
B
B
B
Here
first,
we
we
are
showing
you
what
we
are
going
to.
I
forget
sorry.
I
forgot
the
first
slide
to
introduce
the
demo
and
let
me
get
back
to
the
to
the
side
deck
and
show
you
what
the
environment
of
the
demo
looks
like.
So
the
environment
is
the
the
how
it
looks
at
telecom
italia.
We
have
a
lab
where
that
is
hosting
a
management
stack
available
here
with
a
launcher
in
it.
B
B
B
B
B
B
B
B
B
B
B
B
B
So
what
we
do
for
that
is
updating
the
ml
file
by
saying,
and
especially
the
elm
values
and
saying
that
we
want
three
components
and
then
we
we
apply.
This
in
gate
fleet
will
take
this
into
consideration
and
reapply
the
the
m
chart,
so
that
we
are
going
to
see
that
the
the
amf
a
new
component
is
going
to
get
deployed
on
the
remote
cluster.
B
A
Yeah
thanks
matthew,
so
it
was
very
important
for
us.
We
have
a
strong
interlock
with
a
presentation
of
this
morning
with
swisscom
and
also
from
dt
to
to
perceive
the
transformation
of
the
apparatus
on
a
cloud
native
operator.
So
it
was
the
first
part
and
then
matthew
demonstrates
that
in
a
couple
of
months,
because
we
started
the
technical
stuff
in
december.
So
in
six
months
we
we
are
the
first
demo
on
the
oran
use
cases
here
it's
on
5g
core,
but
we
are
not.
A
We
are
in
the
middle
of
journey
where,
as
I
mentioned,
we
want
to
target
a
production
grade
releasing
so
and
to
to
be
published
also
in
open
source.
So
we
still
need
to
come
back
maybe
next
year
to
to
explain
how
this
project
now
has
managed
the
way
to
publish
it
in
open
source.
We
are
seven
members,
actually
several
industrial
members,
so
how
we
can
integrate
some
new
members
and
we
will
demonstrate
that
we
will
move
the
ecosystem
to
to
achieve
this
transformation.