►
From YouTube: Knative Community Meetup #4
Description
This virtual event is designed for end users, a space for our community to meet, get to know each other, and learn about uses and applications of Knative.
In this meeting, we will hear working group updates. There will be 1 demo: “Building Cloud Native Apps with Knative and TektonCD” by Carlos Santana, senior technical staff at IBM (16:41m - 58:40 m) At the hour (10 am PT), Chris DiBona, director of open source at Google, joined the community meetup to respond to any questions about the Open Usage Commons organization, where he is a board member.
A
A
A
So
without
any
further
ado,
I'm
going
to
go
ahead
and
start.
We
have
a
short
update
about
a
public
steering
committee
meeting.
This
has
been
a
request
that
has
come
up
from
the
community
several
times
and
yesterday
I
heard
that
the
steering
committee
decided
that
there's
gonna
be
a
public
steering
meeting
on
the
third
wednesday
of
each
month,
starting
next
month.
So
august
19,
I
think,
will
be
the
first
public
steering
committee
meeting
that
anybody
can
attend.
D
F
Yeah,
so
there's
so
we
got
the
eventing
bits.
This
includes
inventing
messaging
and
flows
to
v1,
with
the
discovery
api
still
being
defined.
We
did
not
move
the
event
types
forward
because
we
can
say
my
building
in
the
discovery
api,
so
so
that
was
good.
Nice
thanks
for
all
the
work
for
everybody
who
jumped
in
and
let's
see,
we
have
a
few
other
brokers
in
the
work.
F
So
I
know
vmware
is
working
on
on
a
rabbit,
empty
broker,
and
I
know
google
is
working
on
google
broker
and
there's
kafka
broker
work
being
done
by
red
hat
folks.
I
think
it
is
so
I'm
super
excited
to
go
and
see
what
what
comes
out
for
the
next
release.
For
that
also,
the
sourcing
stuff
is
moving
forward
as
well.
It's
gonna
do,
I
know
nacho
and
lionel
have
been
doing
some
work
there
as
well.
So
is
that
two
minutes,
I'm
done
good
work,
everybody
yeah
yeah.
C
I
can
do
a
quick
summary,
so
there's
several
new
features
in
serving
serving
also
has
introduced
a
feature
flag
process
to
enable
people
to
opt
in
early
to
features
that
we
are
testing,
and
I
also
want
to
thank
a
number
of
people
who
end
users
who've
been
contributing
to
serving
by
reporting
issues
with
things
like
high
cpu
usage
or
high
memory
usage
of
various
components,
particularly
at
scale.
These
you
know
end
user
when
the
rubber
meets
the
road.
B
G
Yeah,
so
we
did
a
new
release
there,
so
actually
we
added
some
support
for
broker
management,
which
is
which
is
a
good
thing
and
some
ux
improvements
which
are
which
are
quite
yeah.
I
think
useful
for
people
like
allowing
a
scale
option
which
sets
the
min
scale
and
max
one
with
one
option,
and
yes,
that's
mostly
it.
So
we
also
have
the
opportunity
the
possibility
to
pick
up
our
resource
file
directly
to
create
the
service
which
actually
bridges
the
cni
operations
with
the
resource,
definitions
and
files.
G
C
C
E
E
Yeah,
so
in
the
last
over
the
last,
I
guess
four
or
five
weeks
we
have,
we
have
been
doing
a
retrospective
on
different
facets
of
the
functions
working
group
and
like
outcome
that
that
we
weren't,
I
don't
think
anybody
was
satisfied
with
the
the
outcome
of
the
the
functions
working
group
and
question
of
project
scope.
So
we
have
been.
E
I
guess,
we've
had
three
retros
now
focusing
on
different
layers
of
this
and
the
one
of
one
of
the
things
that
has
received
a
lot
of
attention
and
that
we've,
given
a
lot
of
attention
in
this
retro
process,
has
been
what
should
the
scope
of
k-native
the
github
project,
github
org,
and
and
what
do
we
see
as
like
the
project
scope?
How
will
we
characterize
that
and
how
does
a
new
idea
go
from
being
something?
E
E
I
would
say
that
probably
one
of
the
biggest
takeaways
that
I've
I've
gotten
from
this
this
whole
experience
is
that
people
find
the
idea
of
being
able
to
take
one
of
those
new
ideas
and
and
eventually
have
it,
become
part
of
what
we
think
the
core
scope
of
k-native
is
to
be
really
meaningful
and
that
the
community
wants
there
to
be
such
a
path.
So
all
of
the
recordings
of
those
have
been
posted
onto
the
dev
mailing
list.
E
B
E
H
I
was
it
just
occurred
to
me
that
you,
you
mentioned
the
comment
that
people
find
it
valuable
to
make
a
feature
and
then
have
it
be
promoted
into
what's
considered
core
of
k-native.
I
wonder
if
you've
also
seen
some
sort
of
like
ideas
start
in
k-native
and
get
promoted
into
kubernetes
same
sort
of
path.
E
H
E
Let's
see,
let
me
see
if
I
can
summarize
and
if,
if
I
leave
anything
out,
anybody
can
feel
free
to
jump
in,
so
maybe
I'm
actually
not
up
to
summarizing
this
now,
I'm
reading
it
anyway,
doug
doug
gave
a
pretty
good
treatment.
I
think
that
the
next
details,
really
that
we
have
to
to
figure
out,
are
exactly
what
you
know.
What
kind
of
guidance
could
we
give
around
like
what
are
some
things
that
are
maybe
necessary,
but
not
individually,
sufficient?
E
C
A
Cool
thank
you
for
the
update.
I
think
I
have
to
interrupt
this
conversation
if
that's
happening,
or
maybe
just
kind
of
like
bring
it
to
a
close,
but
I'm
really
happy
that
you
shared
that
update
poll.
I
think
it's
an
important
discussion
for
the
canadian
project,
so
we
can
maybe
agree
to
follow
up
on
that
on
the
next
community
meetup
and
see
where
things
are
at
and
now
I
have
to
pass
the
mic
on
to
carlos
who
is
going
to
do
a
demo
carlos.
B
A
B
B
So
I
what
I
tried
to
do
in
that
in
that
conference
was
consolidate
kind
of
to
the
minimum
demo
possible
that
somebody
can
just
install
k
native
and
tecton
and,
at
the
same
time
use
understand
a
little
bit
of
what
it
is.
So
I'm
not
going
to
do
slides
today.
Just
going
to
talk,
I
have
I
have
slides,
you
can
watch
them.
It's
like
two
slides,
but
the
idea
is.
I
want
to
work
on
on
lowering
the
barrier
to
entry
for
new
commerce.
B
So
what
I
tried
to
do
was
create
something
very
simple
that
is
super
easy
to
install,
so
people
can
actually
see
somebody
else
doing
it
so
because
a
lot
of
times
we
see
demos
and-
and
we
just
see
like
oh-
I
installed
this
software.
Let's
move
on,
let's
explain
the
software,
but
a
lot
of
people
get
stuck.
Would
you
leave
a
lot
of
people
behind
like
well?
How
do
I
install
it
in
my
laptop?
How
do
I
delete
it?
How
how
much
do
I
need?
B
So
that's
what
I'm
going
to
go
over
and-
and
maybe
this
is
a
good
recording
for
folks,
watching
the
recording
so
as
a
message
to
those
that
are
watching
if
you're
not
live
in
the
meetup,
we
got
the
jokes
and
the
bloopers,
so
you're
missing
so
attend
the
next
one.
So
this
is
try
to
do
a
markdown
you
can
with
kubernetes.
This
is
going
to
be
kubernetes.
You
don't
have
to
be
an
expert
granted
or
demo
for
this,
so
you
can
pick.
B
I
have
three,
you
have
karakora,
you
can
go
into
karakoram
just
and
just
run
something,
and
I
don't
know
why
this
is
not
opening.
Let
me
move
this
out
of
the
way
you
can
get
a
free
cluster
in
ibm,
very
it's
very
small.
It's
like
two
gigs
for
cpu,
but
and
then
I'm
going
to
do
the
mini
cube.
Also,
you
can
use
kind
if
you
want
onto
the
time,
but
so
what
I
did.
I
I'm
trying
to
see
also,
for
example,
for
demo
purposes
and
for
educational
purposes.
B
I
want
to
give
a
class.
What
is
the
minimum
amount
of
resources
that
you
need
on
your
laptop
to
run?
These
type
of
things
that
you
just
want
to
learn
so
for
this
example,
I'm
going
to
use
just
serving
and
tecton
so
two
cpus
and
two
gigs
is
good
enough.
So
if
you
have
a
laptop
that
can
do
that,
that's
okay!
I
already
did
so.
B
So
I'm
going
to
use
docker
because
that's
a
common
one,
but
you
can
use
any
registry
out
there
and
set
up
git
just
clone
this
repo.
So
let's
get
started
with
installing
tecton.
So
a
lot
of
people
ask
like
how
difficult
to
install
tecton
there's
an
operator,
but
I'm
going
to
just
install
the
yaml,
and
then
you
run
you
run
these
commands.
Basically,
this
install
the
crds
and
install
a
couple
of
pods.
They
don't
consume
a
lot
of
cpu
and
it
should
not
take
that
long
to
install.
B
So
we
have
good
documentation
in
the
canadian
site,
but
sometimes
I
I
do
this
so
I
can
tell
folks
hey
go
here,
and
this
is
just
one
instance
of
of
getting
something
working
right.
It's
not
the
only
only
way
so
that
took
27
seconds
so
also
to
showcase
some
of
the
improvement
that
people
have
done
around
the
project
of
this
styling
experience.
B
The
next
thing
that
we
need
is
a
courier
courier
is
a
the
networking
part
that
k
native
configure
so
back
in
the
days
was
only
istio.
So
these
days
is
a
blogging
architecture,
so
we
have
courier
istio
contour
glue,
so
all
of
them
are
in
the
in
the
in
the
documentation,
I'm
using
kyrgyz.
It's
it's
tiny
and
it's
what's
easy
to
get
it
get
it
up
all
of
them.
I
guess
they're
easy,
but
this
is
like
I
said
one
example:
I'm
going
to
be
using
no
ports.
B
B
B
So
in
this
case
my
mine
is
called
demo,
I'm
using
the
nip.io,
so
I
already
did
this.
One
thing
that
we
have
to
do
is:
is
patch
tell
kubernetes
about
this
domain
name,
so
I
can
patch
I
can
patch
it.
The
next
thing
is
for
courier
to
work
in
this
environment.
I
need
to
access
some
port
80.,
so
I'm
going
to
create
a
a
like
another
service
that
give
me
a
no
port
on
that
external
iep
and
I'll
be
able
to
access
that.
B
Then
we
have
to
tell
k
native
about.
If
are
you
going
to
use
istio?
Are
you
going
to
use
glue
so
in
this
case
we're
going
to
say
please,
k
native
when
you
want
to
configure
the
networking
correct,
go
ahead
and
create
ingress
resources
for
for
curry
to
pick
up,
and
that's
it-
that's
that's
the
stuff.
B
I
also
wanted
to
show
to
to
users
that,
even
if
you're
watching
it
just
took
me
like
five
minutes
to
have
a
native
install
since
this
is
what's
about
text
on
I
I
also
like
us
knock
in
like
for
k
native,
but
some
people
don't
know
what
k
native
is,
or
they
haven't
seen
a
demo
or
they
cannot
run
a
demo.
So
what
I
did
was
created
a
kind
of
like
a
native
quick
demo
to
showcase
some
of
the
features.
B
So,
for
example,
I
I
like
to
use
the
the
cli
for
when
I'm
like
hacking
things,
so
the
cli
got
updated
to
dot
16.
We
were
just
talking
about
dot
16..
So
all
this
is
running
the
latest
version
of
k
native.
So
I
can
create
a
a
a
k
native
service
with
just
a
container.
I
can
give
it
a
window
of
like
how
fast
do
you
want
to
scale
down,
but
the
cli
give
you
like
a
quick
way
to
get
to
get
started
from
a
container
to
a
url,
for
example.
B
So
in
the
second
part,
we're
going
to
see
text
on
how
we
can
get
source
code
into
a
container,
because
so
that's
the
second
question
that
everybody
asks
is
like:
how
do
I
get
my
code
into
a
container?
So
then
k
native
can
use
it.
So,
since
this
is
a
brand
new
cluster,
I
just
did
a
mini
cube,
delete.
It
doesn't
have
the
the
image
cache.
B
So
that's
why
it
takes
a
little
bit
of
time,
the
first
time,
and
I
think
that
let
me
see
if
I,
if
it
worked
so
if
I
do
this,
I
see
that
is
it
says
that
it's
ready
and
I
get
a
url.
So
if
I
went
from
a
container
registry
having
a
container
to
a
url,
basically
I
don't
know
if
people
are
telling
me-
maybe
I'm
talking
too
much,
and
this
will
be
faster
if
you're
trying
yourself,
but
we
can
just
hit
it
right.
B
I
open
that
that
domain
name
and
it
says
hello
world.
So
let
me
do
it
again
to
make
this
bigger
and
that's
kind
of
like
the
the
aspect
of
it,
but
the
things
that
is
wrong.
The
the
thing
about
k
native
is
on
the
background,
there's
a
but
running
and
when
I'm
not
hitting
it
it's
terminating,
so
I
just
use
a
flag
to
make
it
faster
for
the
demo,
but
the
pot
is
gone.
We
don't
have
any
pots,
there's
no
instance,
so
it's
still
still
terminating
at
that
point.
B
Your
natives
tell
kubernetes
go
ahead,
delete
this
pod
and
it
should
be
gone
in
a
few
seconds,
but
it's
not
going
to
consume
time.
So
that's
the
that's
the
what
we
people,
some
people
call
serverless
right,
the
serverless
aspect
of
k-native
and
that's
kind
of
like
the
quick
demo.
If
you
don't
have
a
lot
of
time
just
to
watch
a
quick
demo,
I
got
it.
B
I
got
k-nav
installed,
so
I
don't
know
if
I'm
going
to
have
a
lot
of
time
here,
but
I
also
want
to
show
another
nifty
thing
that
what
was
playing
with
the
cli
I
was
able
to
to
figure
out
was
what
about.
If
I
want
to
update
something
about
that
service,
so,
for
example,
an
environment
variable,
I
can
do
it
with
the
cli
right.
B
I
can
say
update
my
that
service
and
now
I
have
a
new
revision:
let's
call
it
v1
and
this
one
v1
and
now,
if
I
call
it
my
the
the
app
now,
I'm
calling
the
my
new
revision
of
that
or
that
service.
So
now
it
says
hello,
v1
and
that's
okay,
but
what
about?
If
we
want
to
do
traffic
splitting
I'm
not
using
istio
by
the
way
I'm
using
curves?
So
you
can
get
this
just
with
the
current
gateway.
B
So
that
way,
I'm
able
to
do
traffic
splitting
in
other
in
other
tools.
So
this
is
an
abstraction.
So
so,
if
we
describe
it,
you
will
see
I'm
also
doing
a
demo
of
the
cli.
You
will
see
that
the
information
about
the
revision.
You
will
see
the
75
and
the
25
split
and
the
url.
So
if
I
call
it
the
url,
let's
say
for
in
a
loop,
you
will
see
that
most
of
the
requests
are
going
to
be
going
into
v1
and
then
v2
will
get
a
few.
B
So
this
is
a
common
practice.
When
you're
rolling
out
a
new
software
that
you
roll
out
small
percentage
of
of
the
of
the
traffic
to
that
new
one.
If
it's
loose
bluetooth,
then
you
move
forward.
So,
as
you
can
see,
v1
is
getting
more
more
traffic,
but
what
about?
If
we
want
to
do
something
even
like
beyond
that?
Let's
say
that
I
have
a
a
new
version
of
the
of
my
application.
I'm
going
to
configure
something,
but
I
don't
want
to.
I
don't
know.
I
don't
want
anyone
to
see
it
by
default.
B
So
let's
say
that
I
run
this
command
and
this
what
it's
going
to
do
is
leave
the
75
percent
25
to
v2,
and
then
the
latest
will
get
zero
percent
and
you
will
ask
like
well
to
get
0
percent.
How
do
I
access
access
this
app?
So
this
is
what
is
called
canary
deployment
or
dart
launch.
I
still
don't
want
to
everybody
to
get
it,
but
I'm
going
to
be
able
to
get
it
in
a
second.
B
So
if
we
describe
it
so
describe
this
common
command,
even
in
kubernetes,
to
describe
a
resource.
So
in
this
case
you
see
0
to
the
latest
25
and
20
25
to
v2
75
to
v1,
so
we
call
it
so
it's
not
available
it's
getting
0
but
because
I'm
using
tags,
one
cool
way,
one
cool
thing
that
is
happening
is
k
native,
is
giving
me
a
a
a
domain
name
that
has
the
vitra
v3
tag
in
there.
B
So
that
way
I
can,
I
can
access
it
and
I
can
see
100
of
it.
So
now
I
can
go
to
the
app
and
instead
of
seeing
v1
and
v2,
I
only
see
v3.
That
is
the
latest
one,
and
and
again
this
is
the
flexibility
of
of
k
native,
and
this
is
like
one
of
the
reasons
people
ask
like
if
I'm
using
k
kubernetes
how
what
are
the
things
that
k
name
is
going
to
bring.
So
you
see
some
of
the
abstraction
and
the
fast
way
I
can
get
this
working.
B
I
can
also
do
something
like
this
right.
I
can
say.
Oh
there
was
a
problem
with
v3.
I
think
we
should
go
to
v1,
I
can
say
well,
v1
now
gets
the
traffic,
so
all
the
revisions
are
live
in
the
cluster.
They
live
in
the
cluster,
so
I
can
run
the
loop
again
and
we're
both
rolling
back.
So
this
is
what
we
call
like
rolling
back
to
a
version
that
was
working.
B
So
with
that
the
the
thing
is.
This
is
what
I,
what
I
said
is
if
I
now,
because
I'm
using
tagging,
which
is
very
cool,
I
can
access
any
of
the
versions
and
basically
it's
going
to
bring
up
the
pods
on
demand,
serverless
and
the
available
v1,
v2
and
v3.
But
in
this
case
one
of
them
is
its
latest,
but
they're
always
available
and
you
can
grow
back.
You
can
grow
forward
and
you
can
manage
that
in
other
software
scenarios
this
is
gets
very
complex.
B
So
this
is
what
I
think
the
community
here
in
canada
is
going
to
to
address.
So
some
folks
are
not
familiar
with
it.
It's
hard
to
to
comprehend
between
all
the
noise
or
what
you
get
there
so
and
and
that's
it
so
this
is,
I
just
just-
did
the
demo
with
the
cli,
but
there's
there's
chamo
behind
it.
So
you
can
all
everything
that
I
did.
B
You
can
declare
the
jambo
from
kubernetes
api
resources,
so
I
can
declare
a
v1,
v2
and
v3,
and
I
can
declare
my
split
up
percentages,
my
traffic
there.
So
this
is
a
very
small
version
of
the
yamu.
If
you
were
going
to
do
it
with
like
raw
kubernetes
or
other
tools,
it
would
be
a
lot
of
configuration.
So
that's
what
k
native
is
giving
you.
So
that's
that's
the
idea
and
you
can
deploy
them,
deploy
the
yamaha
I'm
going
to
do
it
here
and
then
you're
back
to
it.
B
B
J
B
B
D
B
Yeah,
that's
that
was
that
was
the
idea
to
start
start
with
the
cli,
but
also
show
that
there's
declarative
way
behind
it.
Right
and
also
I
didn't
want
to
show
all
features
of
k
native
because
then
will
be
becoming
a
long
demo,
and
I
didn't
have
that
much
time
which
I
don't
have
here.
But
I'll
move
to
tekton
next
I'll
go
ahead
and
then.
C
B
B
B
Client,
yes,
so
I
didn't
do
it
one
because
it
will
make
I
this
was
focused
for
that
for
that
presentation,
so
I
didn't
want
to
add
eventing,
also
k
native
cli
one.
I
I
want
to
highlight
the
same
thing
with
the
k
native
cli,
so
I
think
that
the
recent
release
has
the
things
that
I
wanted
to
show
with
the
client
and
show
the
yellow
so
yeah.
I
can.
B
I
can
add
that
and
also
looking
into
the
the
resources,
if
I
add,
k
native
eventing,
I
think
the
dispatcher
is
asked
for
one
more
cpu,
so
I
would
need
like
three
cpus
instead
of
two
cpus,
that's
something
I
also
want
to
go
back
to
the
community.
So
I
have
a
demo
version
or
just
have
one
command
that
that
patches
the
request,
but
I
will
need
more
cpu
and
the
free
cluster
that
I
get
in
iks
is
two
in
ibm.
It's
two
cpu
side,
so
yeah
the
next,
the
the
short
answer.
B
I
B
Yeah
and-
and
this
is
for
demo
purposes
right-
take
into
account
this
is.
I
want
to
run
this
like
even
a
raspberry,
pi
or
vegan
variant
thing.
So,
let's
move
to
tecton,
I
think,
even
like
in
during
the
beginning
of
the
meetup
that
was
asking
about
tecton,
so
how
the
same
concept
of
tecton
like
how
difficult
it
is.
It
takes
a
week.
It
takes
one
day
to
install
text.
Do
I
need
to
buy
a
cluster?
Do
I
need
to
like
put
a
credit
card?
No,
it's
it's
the
same
thing.
B
You
can
start
very
simple
installing
the
same
way
I
did
with
with
k
native.
So
let's
install
tekton,
I'm
still
using
my
two
cpus
right.
I
haven't
done
my
budget
for
my
two
cpu,
so
my
two
gigs
of
memory,
so
I
have
worked
with
folks
that
don't
have
computers
that
have
a
lot
of
memory
so
even
to
try
things
is
a
a
bird
to
entry.
So
that's
why
I
want
to
do
the
demos
with
the
monom
memory
so
for
tekton
you
have
to
install
the
tecton
pipelines
detect
on
this
and
github.
B
They
have
a
project
called
pipelines.
So
you
install
that
one.
It
gives
you
a
web
hook
and
a
controller
also
there's
a
dashboard.
That
is,
I
make
it
optional
some
because
sometimes
I
just
use
the
cli.
I
don't
use
the
the
ui
also
sometimes
I
use
openshift
and
openshift
has
their
own
ui,
but
the
dashboard
is
kind
of
nice
to
have
it.
So,
let's,
let's
put
it
in
there
and
it
also
fits,
doesn't
consume
that
much
memory
to
access
this
ui.
B
I
usually
use
it
not
for
like
configuring
things,
but
just
to
watch
watch
watch
what's
available,
something
people
should
do
with
the
kubernetes
dashboard,
so
the
dashboard
is
there.
So
let
me
create
a
a
service
that
exposed
the
noteport,
so
I
can
access
it.
That
way,
I
can
get
a
url
and
I
can
open
this
url
and
say
it's
a
a
way
to
like
see
what
is
the
pipeline,
what
the
tasks
are
running,
that's
run,
so
I
usually
do
dash
run
fast,
run
or
pipeline
run.
B
So
I
get
I
got
tecton
installed.
So
sometimes
we
spend
because
that's
our
work
as
platform
people
or
tool
people.
We
spend
a
lot
of
time
working
the
stop
process
and
talking
about
the
stop
process
and
some
users
really
like.
I
don't
care,
I
just
need
to
install
it
right.
It
should
work,
so
that's
that's
kind
of
the
net
of
it.
You
install
it
and
you
move
out
of
the
way
I
already
have
the
tekton
cli
sdkn
and,
let's
verify
if
it's
working.
B
This
is
kind
of
a
step
I
put
in
there
to
see
if
everything's
working
so
my
bots
are
working.
I
have
the
dashboard,
I
have
pipelines
and
I
have
the
the
service
so
in
terms
of
people
ask
like
which
cicd
pipeline
do
I
use
for
k
native
well,
keynative
is,
for
the
most
part
runs
of
kubernetes.
Is
you
have
to
build
a
container
so
any
any
tool,
cicd
tool
that
asks
you
to
take
source
code
and
convert
that
source
code
into
a
container?
B
It
works
with
k
native,
if
so
hold
on,
for
the
second
part
on
how
to
do
deploy
the
application
folder.
So
building
for
the
credentials,
I
need
to
I'm
going
to
create
a
the
idea.
Is
I'm
going
to
take
some
source
code
and
push
it
into
the
container
register?
In
this
case
it
could
be
coil
docker
registry.
I
google
registry
amazon
registry,
your
own
private
registry.
I
didn't
register
anyone,
but
then
it
doesn't
matter
cause
tectonic.
B
What
it's
doing
is
taking
the
the
source
code
and
convert
converting
that
into
an
archive
a
container
register
and
then
pushing
it
there.
So
I
already
have
the
secret
created.
I
just
put
my
username
and
password
for
for
docker
hub,
so
I
have
the
secret
there.
The
steps
are
here.
You
can
run
it
and
give
you
the
the
thing
and
then
then
tekton
needs
a
is
a
service
account
because
it
would
run
these
pipelines
or
tasks
the
cicd
jobs
with
a
service
account.
B
So
we
have
to
create
one
and
associate
it
with
this
with
this
secret.
So
because
it's
going
to
be
the
one
that
is
going
to
deploy
and
then
this
step
is
to
configure
the
r
back,
because
later
we're
going
to
use
tekton
to
deploy,
you
like
basically
do
cube
ctl.
So
it
needs
access
to
kubernetes
resources.
Give
it
permission
to
say:
hey.
You
can
create
a
k
native
resource
basically
to
deploy,
so
I'm
going
to
create
that
that's
just
an
art
back.
B
I
created
a
resource,
a
role
base
so
that
way,
if
people
only
have
access
to
one
lane
spay
or
one
project
in
terms
of
openshift
that
only
cluster
raw.
So
that's
another
misconception
that
you
need
like
cluster
admin
for
certain
things.
You
can
just
do
it
in
a
namespace
or
project,
so
the
building
building
tecton
task.
So
a
task
is
the
contact
on
it.
It
will
take
our
source
code.
That
leaves
a
let's
say
in
this
github
repo.
B
I
happen
to
put
both
the
source
code
and
the
pipelines
in
the
same
repo
as
best
practice.
You
will
separate
those,
but
for
demo
purposes
this
is
a
node.js
app,
but
in
case
I
put
a
folder
called
node.js.
You
can
change
the
repo
and
put
here
go
java
python,
any
language
you
want.
So
basically
it's
going
to
take
this
source
code
and
convert
it
to
an
image
in
this
case,
there's
no
imperative
way
of
creating
these.
B
B
B
So
this
task
does
two
things:
it
will
do
a
git
clone
of
the
github
repo
and
then
put
that
in
a
in
a
volume
mount
that
is
empty
directory,
sharing
the
same
thing
and
then
run
a
a
utility,
so
this
utility
could
be
a
binary
like
mechanical,
build
or
something
that
would
take
source
code
and
convert
to
image
like
source
to
image
or
s2i.
But
that's
to
some
developers
that
you
know
that's
not
even
important.
B
I
have
code
in
a
full
code
in
the
photos
you
can
even
use
buildback
right
to
detect
what
type
of
language
you
have
and
the
other
day
it
will
give
you
back
a
container
so
to
deploy
it.
That's
where
you
will
need
cube
ctl,
so
you
will
run
cube,
cto
and
deploy
the
task
after
that,
then
you
can
use
the
tecton
cli
to
list
the
task.
So
let
me
put
this
on
the
top,
so
I'm
just
I'm
listing
the
task.
B
B
Native
build
which
basically,
what
you
were
able
to
configure
declarative
the
source
code
url
and
then
it
will
convert
that
into
an
image,
but
that
project,
I
guess,
matured
and
the
broader
scope
and
then,
if
that
got
converted
into
tekken.
So
if
I
start
this
with
this
command
line
and
I'm
going
to
I'm
specifying
the
the
folder
which
is
node.js
and
the
service
account
which
is
pipeline,
and
this
will
start
a
pot
at
the
end
of
the
day,
this
is
running
pods,
it's
kind
of
like
serverless,
also
serverless,
ci
cd.
B
It
will
launch
a
pod
run
these
two
containers
for
the
git
clone
and
then
convert
the
image
and
pushing
the
image,
and
then
the
bot
will
complete
and
it
will
stay
there,
not
consuming
any
memory
or
cpu.
So,
as
you
can
see,
it
already
did
the
so
I'm
envy
with
them,
because
they
have
colors
and
k
native
cli
doesn't
have
colors.
B
So
that's
just
something
that
I
I
wish
we
had
in
k
native
and
I
I
think,
that's
also
possible
to
add
not
a
problem,
but
so
it's
doing
the
first
task,
which
is
the
geek
clone
and
the
second
one
is
the
one
that
takes
more
time
and
you
can
optimize
this
with
better
way
of
managing
your
docker
files
or
the
tool
that
you're
using
or
caching
into
a
volume
and
such.
But
for
this
demo
I
don't
think
it
takes
more
than
a
minute
to
build.
B
So
I
think
that's
okay,
to
wait
a
minute.
So
basically
already
did
the
docker
build.
So
now
it's
pushing
to
my
container
registry.
So
it's
pushing
to
docker.io
this
namespace,
and
this
is
that
so
this
is
this
is
when
people
ask
about
cicd
is
find
finding
a
tool
or
finding
a
devops
cicd
pipeline
that
will
convert
a
url,
take
source
from
a
url
convert
into
an
image
and
then
push
it.
So
it
got
done
in
one
minute
and
29
seconds.
B
I
can
double
check
if
the
image
is
there
with
a
curl,
because
if
we
want
to
do
it
or
just
talk,
go
to
docker
hub
not
going
to
go
over
that,
so
that's
it
you're
done
with
the
build.
So
this
is
the
part
of
of
tekton
that
is
is
is
you're
writing
scripts
and
you
run
run
them
in
sequence,
and
this
is
one
task
and
it's
reusable.
So
that's
a
reasonable
part
of
tekton.
The
second
part
is:
how
do
I
deploy
now
this
image?
B
I
can
go
back
to
do
a
kn
service
create,
or
I
can
maybe
use
something
like
flux
or
argo.
Those
are
declarative
things
that
would
do
a
git
ups
approach
or
you
can
also
run
like
a
script
using
text.
So
in
this
case
I
have
a
text
on
task
that
basically
takes
a
github
url
that
I
have
the
the
yaml
that
points
to
this
image.
That
is
going
to
point
to
this
image
and
basically
it's
going
to
do
the
same
thing.
Do
a
task
with
two
steps.
B
The
first
one
does
the
kick
clone
of
the
source
code.
They
have.
My
yellow
one
of
the
inputs
is
the
the
directory
where
the
yaml
is
the
the
yaml
file
name
and
basically
does
the
cuba
city
are
applying.
So
this
is
why
I
did
they
are
back.
So
if
you
specify
an
image,
then
I'm
going
to
replace
the
image
in
the
service.yaml
with
the
new
image,
because
on
that's
part
of
the
demo,
the
purpose
is
every
time
you
do
a
commit
to
the
gate
repository.
B
I
want
another
immutable
image
with
a
tag
associated
to
the
hit
command.
So,
let's
deploy
the
deploy
task
again,
keep
cto
apply,
so
I
now
should
have
two
tasks.
I
should
have
a
build
and
a
deploy
task,
so
I
have
the
build
and
deploy
and
I
can
just
describe
the
deploy
again
showing
showcasing
the
detector
cli.
They
put
the
parameters
here.
You
see
the
steps
that
it
takes
and
it
has
a
I
didn't
notice.
B
It
has
a
feet
put
there
and
any
tasks
around
I
haven't
run
a
task
round
yet
so
this
is.
This
is
the
yamo
that
we're
going
to
deploy.
So
basically
you
put
anything
that
you
want
in
those
in
those
scripts.
There's
just
bash
scripts:
keep
city
apply,
so
this
is
the
ammo
that
is
the
k
native
service.
So
you
see
service,
but
it's
it's
under
k
native.
So
this
is
a
k,
n
a
service,
much
shorter
and
we
have
an
image
and
it
has
our
our
environment
variable.
B
So
if
we
start
this,
it
should
start
another
pod
ephemeral,
I'm
passing
the
the
directory
where
my
k
native
is
on
my
service.
So
I
think
I
had
a
link
here
to
that.
I
can't
even
show
that
here
and
you
can
have
one
yama,
multiple
yaml
files,
so
this
is
the
one
that
are
deploying.
So
I'm
deploying
this.
It
happens
to
be
in
the
same
repository
as
I
said,
best
practices
to
have
it
separate,
but
yes,
that
it
took
13
seconds.
B
So
how
fast
does
it
take
to
take
to
deploy
with
tecton
or
deploy?
Canadian
is
just
launching
a
pod
running,
equip
ctl
command
and
these
tests?
The
steps
I
didn't
mention
are
running
in
containers,
so
you
decide
the
container
that
you
want
to
use
so
it
should
have
the
utility.
So
in
this
case,
I'm
using
one
that
I
created
with
cube.
Cdl
and
jq,
and
you
can
use
this
open
source
and
it's
out
there.
B
B
So
I
basically
did
one
task
to
create
the
image
one
task
to
deploy
and
I
should
be
able
to
again
call
my
new
app
and
it
says-
welcome
to
the
1080
meetup,
which
I
updated,
that
in
a
few
minutes
ago
into
the
yamo
and
to
finish
I'm
going
to
talk
about
pipelines.
So
pipelines
is
how
do
I
put
two
tasks
together
to
work
in
power
to
be
executed
in
parallel
or
be
a
security
and
serially
so
yeah?
B
You
declare
this
as
a
yaml
file,
so
you
declare
a
pipeline
that
is
going
to
take
some
input
parameters,
for
example
the
github
url
the
image
that
you're
going
to
send
it.
Maybe
the
image
stack
now
it's
important,
because
every
time
I
want
to
do
a
gig
commit,
I
want
a
new
image
tag
with
that:
git
hash
or
seven,
seven,
first
characters
and
I'm
going
to
specify
the
two
tasks.
B
The
first
task
is
the
build
that
we
just
saw
and
the
second
one
is
the
deploy,
I'm
basically
going
to
run
them
sequentially,
you
can
stitch
them
by
results
or
just
saying,
run
after
build,
so
I
did
run
after
build
and
then
you
deployed
the
pipeline.
So
this
is
where
most
of
the
like
real
projects
happen,
ruin
pipelines.
The
tasks
are
like
composable
things,
so
it's
a
composable
framework.
B
B
Basically,
I'm
passing
the
the
image
url
where
I
want
to
deploy
and
the
repo
url,
which
is
my
my
repo,
and
it's
going
to
run
the
two
tasks
instead
of
I
did
it
manually
before
it's
going
to
run
one
after
the
other,
using
the
same
image
and
and
then
again
it
will
take
a
minute
to
do
that
and
then
we'll
be
able
to
see
our
k
native
application
and
then,
lastly,
is
the
automation
because
it
gets
tired
by
running
this
thing
manually.
Every
time
you
push
something
to
git
or
you
submit
a
pr.
B
So
if
you
install
it,
let's
see,
if
that's
that's
finished,
I'm
going
to
skip
something
a
lot
of
time
over
over
my
budget.
So,
let's
give
it
a
time
for
this
to
finish,
and
I
will
talk
a
little
bit
more
about
about
the
triggers,
so
triggers
is
a
way
that
you
can
bind
when
you
receive
a
git,
a
git
request.
You
have
a
template,
so
this
is
the
template
is
what
are
the
things
that
you
want
to
create
resources?
B
Oh
so
it's
done
so
it
took
a
minute
and
a
half
so
now
is
is
updated
and
I
was
able
to
create
the
pipeline.
So
let
me
install
the
tekton
triggers
if
you
install
the
tecton
triggers
this
is
a
separate
component,
so
you
can
decide
if
you
want
it
or
not.
So
if
you
need
a
trigger
the
pipelines
with
a
with
a
url
endpoints,
then
you
can
do
that.
So
that's
the
tecton
triggers
and
and
basically
what
you
need
is
a
theory
template
that
says.
B
B
So
this
is
a
nice
thing
that
tectum
has
so
I'm
gonna
be
going
to
replace
it,
because
I
want
to
say
what
is
the
image
that
I
wanted
to
use
and
then
it
has
a
binding
and
then
the
second
component
is
like
when
I
receive
a
webhook,
how
do
I
extract
things
from
the
http
vario
of
that
git
web
hook?
So
I
can
get
get
the
commit
id
in
this
case
is.
This
is
a
format
for
github
the
url,
but
also
it
has
an
extension
that
I
can
going
to
show
of
how.
B
How
can
I
use
text
tecton
to
tell
that
hey
run
some
logic
in
their
declaratives?
So
let's
do
the
binding.
So
this
is
what
is
called
tecton
binding
and
then
the
event
trigger.
This
is
basically
putting
that
bot
in
place.
Here's
the
yaml
again
this
a
lot
of
yammo,
but
basically
what
you're
saying
is.
I
want
a
a
service
or
a
pod
to
spin
up
and
it's
going
to
process.
B
Github
events
in
this
case
push
or
I
can
say
which
one's
one
one
with
a
with
a
cell,
which
is
this,
is
a
a
way
of
declarative
way
of
doing
some
logic
in
here
and
that
is
coming
from
master.
So
I
can
scope
it
to
push
and
master.
I
can
also
take
the
gig
commit
and
just
truncate
it
to
the
seven
characters.
So
that
would
be
my
image
tag.
So
every
time
there's
a
github
have
git
coming
happens.
I
want
a
new
image
with
a
new
gap.
B
B
B
So
in
this
case
I'm
going
to
expose
it
to
renault
port,
so
I
can
access
it
and
then
I'm
going
to
get
the
github
url
like.
If,
if
github
can
access
my
laptop,
it
will
work,
but
I
you
can
use
a
kubernetes
cluster,
free
or
paid
that
is
on
the
cloud
or
you
can
use
inlets.
I
point
out.
Inlets
is
something
that
you
can
like
get
traffic
from
the
internet
into
your
laptop.
B
So
I'm
not
going
to
show
how
to
add
this
in
in
here,
because
I
already
did
it,
but
you
can
test
it
with
coral,
so
you
can
like
emulate
that
github
events.
So
I
can
like
it's
a
push,
so
I
just
triggered
the
pipeline
and
it's
running
and
actually
I
can
even
see
it
in
the
in
the
text
on
here
when
when
it's
running,
let
me
see
this
one
is,
I
think
it's
this
one.
B
This
one
is-
and
I
guess
so-
it's
running
in
the
ui-
so
that's
that's
triggered
in
that
I
can.
I
can
also
show
where
already
did
it
in
the
github
url.
So
if
you're
doing
this
lab,
basically
follow
the
instructions
and
go
into
settings
and
you
click
web
hooks
and
you
can
do
this
programmatically.
Also,
I
added
I
added
one
to
a
free,
kubernetes
cluster
that
I
have
online
in
ibm.
So
I
have
a.
I
have
a
second
like
environment
like
this,
because
I
wanted
to
test
online.
B
So
basically
you
just
edit
this
and
put
the
github
url
either
using
inlets
to
your
mac
or
your
kubernetes
cluster
and
every
time
something
happens
in
git.
I
can
change
something
that
will
start
a
new
change,
so
I
can
go
ahead
and
change
something
about
the
the
application.
So
I
can
say
something
like
changing
the
code,
so
now
it
becomes.
I
only
focus
on
my
source
code.
B
My
java
python
go
code,
and
I
can
put
here
thanks
for
watching
the
demo
video
right
and
this
would
trigger
I'm
going
to
just
commit
to
master
like
yellow
right,
and
this
would
trigger
the
pipeline
in
this
case
I'm
pointing
to
my
other
cluster.
So
it
started
building
my
container.
It's
going
to
deploy
the
app
in
that
name
space
and
then
my
new
app
should
be
available
in
one
minute.
B
So
I
can
do
this
for
prs
and
to
add
serpent
name
space
or
when
I
merge
to
master
to
deploy
to
staging
or
production.
So
I
hope
this.
This
demo
is
useful
for
folks
getting
started
into
this
space
that
it
feels
a
little
bit
intimidated
right,
that
you
need
a
lot
of
resources
or
powerful
computer
or
wallets.
It's
super
easy
to
get
started
and
and
follow
along.
So
I
think
I'm
over
my
30
minutes
mariah.
B
A
A
A
Okay,
yes,
we
can
hear
you-
and
this
is
the
section
where
we,
if
anybody
has
any
questions
about
open
usage,
commons
organization,
I
would
like
to
ask
chris
about
trademarks
for
canadia,
etc.
This
is
the
moment
to
do
so.
There's
there's
a
space
in
the
agenda
where
I
invited
the
community
to
post
any
questions.
This
is
at
the
end,
so
we
can
start
there
or
we
can
start
chris.
If
you
want
to
start
with
anything
else,
feel
free
to
do
so
as
well.
L
L
But
if
you
can
hear
me
well,
I'm
just
going
to
get
started
and
then
you
know
four
or
five
minutes
from
now
I'll
start
walking
down
the
questions
that
people
have
been
in
the
dock,
which
is
great,
I
will
say,
there's
one,
that's
that's
kind
of
baffling
to
me
and
I
might
ask
the
anonymous
tiger
cat
dog
bird
or
whatever,
to
to
help
me
with
that
one.
But
yeah.
L
L
When
I
first
got
here,
we
had
had
some
patching
activity,
starting
in
1999
into
the
linux
kernel
and
some
other.
You
know
haphazard
contributions,
but
we
we
started
a
path
of
formalization,
initially
releasing
four
projects
and
then
now
we
release
between
eight
and
fifteen
projects
a
day
into
open
source
software
under
real
open
source
licenses,
alex
go
ahead
and
put
it
in
the
in
the
docs,
so
I
can
get
to
it.
Otherwise,
I'm
going
to
be
useless
to
you.
L
Our
7
day,
active
and
30
day
active
in
terms
of
getting
contributions,
updates
public
requests.
That
sort
of
thing,
I
don't
need
to
tell
any
of
you
what
it
means
to
be
an
active
project
and
so
of
those
we've
also
donated
some
to
foundations.
So
we've
donated
lock,
stock
and
barrel
trademark
copyright
patents
to
about
12
or
13
projects
of
the
3000
30
day,
actives
into
foundations.
L
Where
we
work
on
projects
like
kubernetes
grpc,
you
know
amp
actually
went
into
a
foundation
recently
and
then
things
like
gosh
wasm,
you
know,
I
think,
would
be
relevant
to
your
community.
They
went
to
the
w3c
where
we've
been
a
member
since
2006
and
the
rest
so
yeah.
So
what
this
all
means,
I'm
not
trying
to
you
know,
put
on
errors
or
stand
up
on
some
high
mountain
saying
how
look
at
how
great
we
are.
That's,
that's
not
the
point
of
this.
L
The
point
of
this
is
we
end
up
hitting
every
problem
intellectual
property
kind
of
before
any
other
company.
Does.
If
you
look
at
you
know,
the
various
lawsuits
brought
against
google
run
intellectual
property
that
have
an
open
source
component,
whether
it
was
around
courgette
and
the
chrome,
auto
updaters
or
the
rather
prominent
oracle
lawsuit.
L
These
are
you
know,
legion
in
the
company,
and
so
we've
also
been
getting
a
lot
of
questions
about
how
people
can
use
those
trademarks
in
commerce
and
in
their
distributions
of
these
softwares.
So,
whether
it's
something
like
istio,
angular
or
garrett,
they
all
have
had
people
come
to
say,
hey,
you
know
we.
We
were
using
this
and
we're
shipping
to
third
parties
we're
selling
services
around
them.
You
know
how?
How
can
we
use
the
the
the
trademark,
the
logo,
the
name
but
the
the
mascot
in
in
many
cases?
L
And
while
we
sort
of
counted
on
the
apache
license
to
express
our
intent
as
a
company
around
non-copyright
non-patent,
you
know
intellectual
property,
like
you
saw
on
sitemaps,
like
you
saw
in
in
a
variety
of
hardware
related
and
and
trademark
related
intellectual
property.
It's
the
fact
of
the
matter
is
open.
Source
licenses
simply
don't
cover
trademarks.
The
apache
license
is
the
closest
one
that
comes
close
to
acknowledging
the
existence
of
trademarks,
and
it
says
specifically
that
it
doesn't
deal
with
them.
L
So
you
end
up
having
to
come
back
to
the
trademark
holder,
in
this
case
google,
for
a
lot
of
folks.
So
what
we
decided
to
do
is
create
a
new
organization
called
the
open
usage
commons.
I
was
originally
thinking
about
calling
it
the
corner
case
commons,
but
nobody
would
get
the
joke
and
yeah.
So
we
created
the
open
usage
commons.
L
We
we
just
literally
just
started
this
thing
with
with
charles
isbold
who's,
the
csd
in
the
georgia
tech
cliff
lampe,
who
is
over
university
of
michigan
allison
randall,
who
I
think
some
of
you
must
know
from
her
travels
in
open
source
and
miles
ward,
who's,
sort
of
our
voice
of
the
customer
on
the
board
of
directors.
So
there's
six
people,
two
googlers
myself
and
jen
phillips
and
then
four
non-googlers
we
all
have
eagle
votes.
So
it's
not
some.
You
know,
stock
puppet
organization
like
some
others
will.
E
L
So
we
think
that
the
board
of
directors
will
probably
grow
over
time.
You
know
we're
just
still
in
the
setting
of
the
bylaws
part
of
the
organization.
You
know
the
reality
of
being
at
google
is
when
you
start
something
externally,
you
have
to
manage,
you
know
talking
to
reporters
and
all
the
rest,
but
we're
just
getting
started.
You
know
one
thing
that
you
know
we
we
think
is
relevant.
L
So
you
know
and
then
for
individual
marks
that
have
demands
beyond
that
we
would
have
an
escalation
committee
that
would
be
staffed
from
the
steering
committees
or
from
the
you
know,
the
league
committers
of
those
projects
you
know
to
handle
escalations,
so
that's
another
place
where
I
think
we
would
see
vendors,
you
know,
but
I
should
point
out.
This
is
not
a
vendor
organization.
This
is
not
designed
to
replace
a
linux
foundation
or
a
cncf.
L
You
know,
which
is
a
benevolent
corporation,
designed
sort
of
like
a
change
chamber
of
commerce
to
represent
its
members.
You
know
we're
specifically
trying
to
create
guidance
for
any
project
to
use
whether
they're
in
the
ouc
or
not
around
how
to
you
know,
maintain
trademarks
consistent
with
the
osd.
E
I
mean
it
it
sort
of
doesn't
the
reason
that
I
ask
about
the
board
of
directors.
Is
that
if
we
really,
if,
if
you
want
ouc
to
have
the
perception
of
neutrality,
I
would
imagine
that
it's
important
for
anybody
to
be
able
to
have
a
seat
at
the
table
at
the
board
of
directors
similar
to
how
okay
you
know
you
can
join
the
linux
foundation,
etc.
That's
that's
sort
of
so
you
can.
L
Join
the
linux
foundation,
but
you're
not
in
charge
of
the
board
of
directors.
Just
by
joining
I
mean
so
listen
the
linux
foundation
is
the
linux
foundation.
The
only
thing
that's
going
to
prove
neutrality
of
the
board
of
directors
of
the
ouc
will
be
time.
You
know
the
people
who
aren't
going
to
trust
us,
no
matter
what
are
not
going
to
come
to
trust
us
a
couple
years
from
now.
L
You
know
so
the
reason
we
compose
the
board
in
the
fashion
we
did
is
that
we
wanted
people
who
had
the
best
interests
of
computer
science
and
open
source
at
heart.
You
know,
and
you
know
people
can
either
trust
us
or
not.
I
I
kind
of
don't
care
you
know,
so
the
real
question
is:
is
there
a
relevant
place
for
other
vendors
in
that
rubric?
And
I
think
the
answer
is
yes,
and
I
think
that
you
know
asked
me
in
three
to
six
months
when
we
got
more
road
under
our
feet.
L
Okay,
anyways,
I'm
happy
to
go
to
questions
now.
The
other
thing
I
would
like
to
mention
is
you
know.
The
neutrality
of
the
organization
is
really
paramount
for
people
to
understand,
so
not
only
did
we
make
it
so
that
you
know,
google
has
only
one
third
of
the
votes
of
the
board
of
directors,
but
also
you
know,
google
can
never
ever
own
more
than
49
of
an
external
body
like
this,
without
invoking
all
kinds
of
rules
and
regulations
around
our
anti-trust
posture.
L
So
it
was
really
important
to
us
that
not
only
are
we
not
owning
a
majority
share
of
this
thing,
we're
also
not
voting
even
more
than
a
third
in
the
current
structure.
So
you
know
neutrality
is
a
legal
thing
for
me,
that's
really
important.
It
doesn't
really
matter.
If
people
like
it
or
not,
you
know
it's.
The
reality
is
someone
like
charles
is
bull
and
cliff.
L
Okay,
relation
of
that's
not
really
a
question.
Oec
is
open
usage,
yeah
sure
relationship
with
the
osc
in
regards
to
the
cncf.
If
k
native
was
in
the
cncf
volcanic
is
not
in
the
cncf,
so
I
don't
know
really
how
to
answer
that.
The
cncf
has
been
pretty
hostile
to
the
creation
of
the
open
usage
commons.
L
So
you
know
whatever
who
actually
owns
the
sql
trademark.
So
the
issue
trademark
is
an
in-progress
trademark
in
the
u.s
patent
and
trademark
office.
So
you
know
from
google
and
any
of
the
common
mark
laws
would
apply
to
it,
so
the
strings
all
over
your
code
base
could
can
even
be
con
infringing.
I
wouldn't
think
so
specifically
when
I
think
about
how
we
will
handle
the
sgo
trademark
again
consistent
with
the
open
source
definition.
L
I
mean
our
main
goal
honestly,
is
that
any
project
inside
the
ouc
will
never
be
considered
non-free
from
the
debian
perspective,
so
we
feel
that
that's
like
the
higher
bar
to
get
through
so
obviously
referencing
istio
inside
the
candidate
of
code
base
would
be
completely
consistent
with
open
source
with
the
open
source
definition
and
the
debian
free
software
guidelines.
So
I
wouldn't
worry
about
that
in
the
slightest.
L
Okay,
yeah
yeah
we're
not
doing
some
game
around,
like
licensing
from
google
to
the
ouc,
it's
a
full
transfer.
So
it's
you
know
it.
Otherwise
it
wouldn't
be
real.
You
know
so
so
the
lf
said
that
under
the
u.s
trademark
law,
you
are
not
able
to
effectively
separate
your
project
mark
from
control
the
underlying
open
source
project.
So
that's
actually
not
completely
true.
L
You
know.
So
if
you
look
at
most
of
our
projects
that
we've
released,
there
is
no
complicated
governance
structure
like
canada
and
istio
and
and
projects
like
kubernetes.
Have
you
know
most
projects
that
we
have
are
are
almost
single
contributor,
if
you
think
about
it,
and
if
you
look
across
open
source
in
general,
most
projects
don't
have
complicated
structures
like
this,
but
yeah.
L
Let's,
you
know,
why
is
trademarking
issues
suddenly
right
now
versus
just
open
project
governance?
I
I
think
that
this
is
orthogonal
question,
because
I
think
open
project
governance
is
absolutely
a
question
all
the
time
for
all
projects.
So
I
think
the
trademark
is
an
issue
for
us
for
a
while,
as
I
mentioned
earlier
in
my
my
intro
there,
so
I
I
think
that
the
question
might
be.
I'm
not
sure.
I
really
know
how
to
answer
this
question.
So
trademark's
been
an
issue
for
us
for
a
while.
L
You
know,
we've
been
hitting
problems
with
it.
We've
been
hitting
difficulty
getting
people
to
accept
when
we,
when
we
issue
trademark
guidelines
on
a
project
by
project
basis,
so
we're
just
trying
to
you
know
have
like
a
really
good.
You
know
osd
compliant
starting
point
for
trademark,
so
yeah
as
far
as
open
project
governance.
That's
that's!
You
know
something
I'm
going
to
push
back
on
to
the
k
native
steering
committee.
They
they
they
do
their
thing.
I
mean,
I
think
dan
cerulli
when
he
was
talking
about
istio's
project
governance.
L
You
know
he
said
it
really
well
on
this
there
was
like
a
podcast.
Maybe
we'd
get
a
link
pasted
it.
Not
many
people
seem
to
be
subscribed
to
that
person,
but
and
since
katie
was
not
one
of
these
projects
going
to
the
you
see
you're
talking
about
this
because
google
plans
to
it
to
go
there,
one
of
the
things
I've
said
over
and
over
again
to
the
canada
folks
is,
you
know,
talk
to
me
in
three
to
six
months.
L
If
ouc
is
actually
useful
and
and
is
making
a
good
impact,
then
absolutely
we'll
we'll
we'll
talk
about
canada
and
and
at
that
point
I
would
hope
the
k-native
community
would
see
it
as
a
as
a
net
win,
as
opposed
to
the
one-off
nature
of
the
trademark
guidelines
around
k-native.
Currently
so
because
the
choice
right
now
is
whatever
plan.
Whatever
google's
giving
to
you
is
what
the
guidance
is
and
so
will
ouc
give
guidance,
that's
better
for
the
commercial
and
open
source,
implementers
and
distributors
of
canadians.
L
I
I
would,
I
would
think
so,
but
you
got
to
let
me
you
know
we
again,
we
have
just
gotten
started.
Let
us
get
our
guidelines
out
there
and
and
through
the
the
ringer
of
the
debian
free
software
you
know,
process
and
and
and
let
people
bang
on
it
for
a
while
until
we
we're
sure
so
given.
Oh,
I
haven't
followed
this
link.
It
feels
like
google
saying
they're.
Okay,
with
mark
ownership
being
lost
to
an
ambiguous
usage.
You
know,
I'm
not
gonna
have
time
to
probably
analyze
a
github
thing.
L
I
I
don't
know
that
yeah.
This
is
a
lot
to
go
through
sorry,
this.
This
was
not
in
the
in
the
briefing
document
up
until
recently.
So
I
don't
really
know
what
to
say.
Google
is
not
saying
that
I'm
okay
with
mark
ownership
being
lost
to
and
doing
ambiguous
usage.
So
there
you
go
unless
we
have
so
yeah
steering
owns
all
aspects
of
google
and
sparks
enforces
usage
of
the
mark
delegates.
So
there's
lawyers
who
declare
whether
there's
a
lot
going
on
here,
I'm
just
gonna.
Let
this
one
go.
L
Sorry,
it's
just
there's
too
much
to
unpack.
Did
the
projects
being
donated,
I
tell
you,
see,
choose
it.
If
I
choose
that,
I
mean
choose
to
donate
their
trademarks
to
to
ouc
here's
the
thing
I
I
don't
want
to
sound
too
autocratic
here,
but
there
are
some
decisions
that
I
make
for
google,
and
these
three
projects
have
a
lot
in
common
from
a
trademark
perspective.
L
They
have
third-party
implementers
shipping
them
and
who
want
to
use
the
words
garrett,
angular
or
istio
when
they
ship
them
when
they
ship
them
to
their
customers
when
they
sell
services
around
them
and
also
all
three
of
them
are
under
intensely
active
development.
I
mean
garrett,
lesso,
garrett's,
very
mature,
but
angular
and
seo
are
under
intense
development
inside
the
company
and
they
were
getting
more
and
more
people
asking
them.
Well,
I
want
to
use
it
this
way.
I
want
to
use
a
little
bit
this
way.
I
want
to
use
the
name
that
way.
L
I
want
to
use
the
mascot
this
way
and
we
wanted
to
stop
having
to
do
one-off
answers.
So
one
of
the
things
about
google
is,
we
do
so
much
in
open
source.
We
do
so
much
with
open
source
that
anytime,
we
start
getting
a
question
more
than
once.
We
need
to
optimize
it.
Otherwise
we're
going
to
be
spending
all
of
our
time.
In
these
little
corner
cases.
E
L
L
L
L
Yeah
I
mean
I
just
need
to
push
back
on
that
on
on
to
the
team.
So
does
google
make
a
unilateral
decision
around
donation
of
k-native,
I'm
assuming
this
is
around
the
cncf
non-donation
of
k-native
that
we
told
you
about
seven
months
ago
or
will
they
allow
the
string
committee
determined
by
the
community
without
a
vendor
majority
making?
So
here's
the
cute
thing
people
keep
on
saying
without
a
vendor
majority
because
they
say
community
and
it's
like
what
do
you
mean
by
community
they're
like
well?
L
You
know
the
plurality
of
the
people
who
contribute
to
the
project
and
it's
like
okay.
So
when
you
know
in
the
last
three
months
in
the
last
year
and
last
two
years
since
the
project
started,
you
know,
do
we
intend
to
make
a
united
decision
around
donated
donation
of
candida
to
cncf?
We
already
did
so.
I
know
a
lot
of
people
didn't
like
that,
but
you
know
that's.
That
is
what
it
was.
L
L
L
E
That
that's
true!
Actually,
chris,
you
I'm
glad
you
brought
that
up,
because
we
we
have
been
working
on
refining
or
or
determining
actually,
since
we're
still
in
the
sort
of
by
fiat
bootstrap
phase
of
the
steering
committee,
how
the
steering
committee
will
be
composed
and
brenda.
I
think
you
have
drafted
a
document
with
some
different
options
that
we
might
look
at,
and
it's
shared
with
k-native
dev
now
is
that
right.
N
E
Yeah-
and
I
think
maybe
like
the
suggestion
in
that
doc
that
you
were
reading
chris
got
maybe
edited
by
multiple
people.
So
so
I
was
I
was
wondering
like
in
in
the
future.
Like
does
google
intend
for
the
steering
committee
and
k
native
to
make
a
decision
about
like
potentially
homing
the
project
in
ouc
or
like?
Is
that
a
google
decision.
L
So
it's
worth
pointing
out
that
the
ouc
does
not
do
project
governance.
It
doesn't
do
project
marketing,
it
doesn't
do
conferences.
All
it
does
is
put
together
trademark
guidelines.
So
you
know
when
you
talk
about
homing
the
project
inside
the
ouc.
That's
that's
just
not
a
thing
right
so
now
owning
the
project's
trademark
in
the
iec.
It
might
be
a
better
choice
than
having
google
hold
it
and
and
decide
what
to
do
with
it
on
a
one-off
basis,
depending
on
the
the
sort
of
the
weather
right.
L
Okay,
is
google
committed
to
a
steering
which
has
no
single
event,
there's
a
lot
of
discussions
here
about
candidate
of
steering
politics
that
I
just
have
nothing
to
tell
you
about.
What's
the
benefit
of
the
project
owner
to
do
this,
since
they
lose
ownership
of
their
trademark
name,
logo
right,
but
then
still
own
the
code?
What's
the
benefit
of
the
split
of
ownership,
I'm
reminded
of
the
founder
movie,
where
all
he
wanted
and
all
that
matters
was
I
don't.
L
I
haven't
seen
the
movie
so
sorry,
I'm
sure
there's
something
about
ray
kroc
that
we
can
draw
from
yeah
so
since
they
lose
ownership,
what's
the
benefit
of
the
project.
So
in
the
case
of
the
three
projects
we
picked,
the
benefit
is,
strictly
speaking,
that
they
are
gonna,
have
proper
guidance
in
accordance
with
the
open
source
definition.
So
it
takes
away
the
ambiguity
around
google's
current.
L
You
know
one-off
process
for
giving
trademark
permission
or
not
permission
enough
depending
on
whoever
you
might
be,
and
so
I
believe
that
if
we
can
give
guidance
consistent
with
the
open
source
definition,
then
the
commercial
concerns
about
if
they
can
ship
and
use
the
word
istio
or
use
the
words
angular
or
whatever
you
know,
which
should
go
away
or
at
least
will
be
a
lot
more
clear.
So
you
know
those
who
are
really
interested
in
just
proprietization
and
the
rest.
M
Just
but
just
just
to
follow
that,
though,
when
it
comes
to
making
a
decision
in
terms
of
whether
somebody
can
or
can't
use,
for
example,
the
logo
in
a
particular
way.
Who
makes
that
decision?
Is
it
the
ouc
board,
or
is
it
the
project
owners
reading
the
guidance
from
the
ouc
board,
who
actually
can
make
those
decisions
so.
L
For
the
projects
that
are
actually
in
the
ouc
whose
trademarks
are
actually
in
the
ouc,
the
the
board
will
be
given
will
produce
guidance
that
can
be
used
and
then,
in
the
case
of
a
really
contentious
issue,
the
idea
would
be
that
there
would
be
someone
from
the
steering
committee
of
that
project
or
from
the
governance
of
that
project.
Who
would
represent
those
more
complex
needs
beyond
the
osd?
So
a
good
example
of
this
would
be
so
suppose.
L
If,
if
you
think
of
k-native,
it's
not
actually
a
software
project,
it's
just
a
list
of
specifications
and
apis
okay.
What
does
it
mean
for
somebody
who
otherwise
doesn't
use
a
single
line
of
code
from
the
k
native
project,
but
presents
an
exact?
You
know
100
percent
complete
copy
of
it?
Can
they
call
themselves
candid?
L
You
know
you
know
these
are
the
kinds
of
questions
we
get
around
like
istio,
for
instance,
and
so
what's
the
right
answer
from
an
osd
perspective,
so
I'm
very
apache
license
oriented.
So
everything
for
me
is
you
get
a
grant
to
copyright,
you
get
a
grant
to
patents
and
you
would
get
a
granted
trademark
if
you
use
the
software.
So
if
you're
using
the
software,
then
you
can
say
you're
using
the
software
right.
L
If
you
decide
to
fork
and
go
another
direction,
is
the
right
language
that
you've
been
derived
from
that
software
or
based
on?
We
don't
really
know
what
the
right
one
is
yet,
but
we
know
there's
something
like
that
in
there.
If
you
create
something
that
is
100
compatible
but
doesn't
use
a
single
line
of
code
from
that
project,
can
you
say
that
you
are
100
compatible
with
angular
or
istio,
or
something
there's,
probably
some
languages?
L
That
is
consistent
with
the
w3
software
guidelines,
but
I'd
imagine
the
answer
would
be
largely
yeah,
probably
not
unless
you're
actually
using
the
software.
You
can't
use
the
name
right
because
one
of
the
things
we
want
to
avoid
are
people
saying:
hey,
I'm
the
I'm
the
istio
project
when
in
fact
they
have
not
a
single
line
of
code
in
comma,
so
yeah.
M
L
L
That
we
would
be
able
to
not
have
to
make
any
decisions
unless
they
fall
under
some
weird
corner
case
in
the
far
side
of
the
open
source
definition,
because
I
I
actually
truly
believe
the
open
source
definition
can't
give
proper
guidance
on
the
use
of
a
name
and
a
logo
and
a
mark-
and
you
know-
and
I
guess
that's
why
I'm
the
open
source
person
at
google
but
yeah
so
so
yeah
and
then
the
more
complicated
questions
of
api
compatibility.
But
you
know
right
now
we're
just
handling
trademarks
as
our
intellectual
property.
L
How
could
one
begin
to
approach
ensuring
a
technically
conform,
a
non-mark,
infringing
downstream
implementation
derived
from
the
uc
project,
because
I
used
to
have
a
roadmap
for
how
its
projects
people,
let
out
so
we've
only
just
started
talking
about
what
conformance
testing
would
mean
with
regard
to
you
know,
is
that
de
facto
permission
to
use
a
mark
again
we're
just
getting
started
I'd
say
come
back
and
talk
to
us
in
a
bit.
You
know
we
need
to
need
a
bit
of
time.
Who
did
google
work
with
for
making
ouc
happen?
L
L
That
way-
and
I
I'll
be
frank
for
you-
I
thought
a
lot
of
the
guidance
that
came
out
of
the
company
was
mixed
and,
and
sometimes
I
felt
it
may
be-
even
an
additional
restriction
on
top
of
the
gpl
or
the
or
the
apache
license,
which
is
something
that
I've
always
tried
to
avoid
at
google.
So
I
was
like
well,
we
need
to
set
up
so
back
in
the
early
days
of
google.
You
know
I
was
very
adamant
that
we
would
not
create
an
open
source
license.
L
L
You
know
when
you're
writing
kernel
drivers
or
whatever
you
use
the
gpl,
we
weren't
using
some
google,
you
know
commercial
source
license
or
one
of
these
new
source
licenses
you've
been
seeing
and
there
isn't
anything
like
that
for
trademark,
and
so
we
decided
that
we
were
going
to
create
something
so
that
we
could
create
that
license
for
lack
of
a
better
term.
That
would
be
consistent
with
the
osd
for
trademark,
and
that
was
the
ouc.
L
So
that
was
something
that
we
had
been
talking
about
inside
the
company
for
about
oh
gosh
for
about
a
year
now,
if
it
was
going
to
be
the
right
approach,
and
so
that's
what
we
decided
to
do,
who
decides
who's
on
the
oec
board?
In
the
beginning,
we
just
went
and
talked
to
some
people
who
we
we
got
along
with
in
the
industry
and
asked
if
they
want
to
join
some
people
said
no
and
other
people
said
sure
yeah.
That
sounds
like
a
lot
of
fun.
L
So
that's
what
we
did
reminds
me
of
docker
renaming
to
moby
for
trademark
purposes.
That
did
not
work
out
too
well
for
docker
popularity
and
trust
dockers.
Populating
trust
is
not
my
problem,
so
I'm
going
to
skip
that
since
iec
must
own
the
trademark,
after
other
foundations,
also
expect
to
own
the
trademark.
E
L
L
Things
about
being
google
is
you
can't
do
anything
without
everyone
really
needing
to
know
everything
about
it,
and
it's,
like
you
know,
we're
in
the
middle
of
working
on
it
right
now.
So
I
I'd
love
to
tell
you
that
we're
going
to
get
it
all
right,
first
pass,
but
there's
no
way
so
we're
going
to
need
people
to
comment
on
it
and
tell
us
what's
what's
right,
what's
wrong,
so
I'd
really
appreciate
that.
E
L
Okay,
I
would
measure
success
specifically
if
people
take
a
look
at
the
guidance
that
we
produce
and
start
using
it
like
they
say:
hey
we've
got
a
trademark,
we're
just
going
to
use
that
stuff,
because
it's
already
been
figured
out,
they've
google's
already
taken
it
on
the
chin
and
figured
out
good
guidance
around
these
things.
So
I'm
just
going
to
use
that
you
know
that
would
be
a
great
outcome
so
that
that
would.
I
A
A
Thank
you.
Thank
you,
everybody
for
joining
and
staying
this
long
and
having
this
conversation,
I
think
it's
important
to
to
have
a
good
understanding
of
what
the
open
usage
commons
organization
is
doing.
Please
make
sure
that
you
fill
out
the
survey
for
this
meetup
and
give
us
feedback
on
the
on
the
content
and
the
event.
Overall.