►
From YouTube: PetroChina: Big Oil’s Information System on Cloud Native Technology – Hui Chi & Kai Chen
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
PetroChina: Big Oil’s Information System on Cloud Native Technology – Hui Chi, PetroChina & Kai Chen, Alauda
The IT capability of General Administration of Customs of the People’s Republic of China (GACC) has been at the forefront of ministries and commissions under Chinese government. In order to cope with the rapid growth of new trade business, GACC has decided to use Kubernetes to build a new generation of cloud platform to create a sustainable technology eco-system for China Customs. How does Cloud Native technology and DevOps help GACC achieve the following goals rapidly? - Build the app platform; - Iterate the intelligent information system rapidly; - Standardize the management system steadily; - Make the workflow standard; - Visualize the operation system; - Intelligentize the fault analysis; -Automate the maintenance operations; …… Maybe you think the e-government system in China has been mysterious, but you can get the answers to all these questions in this speech.
To learn more click here: https://sched.co/G5My
A
Hey
fouling
it
so
actually
on
Tuesday
I,
we
talked
about
how
we
did
DevOps
in
our
team,
because
that's
something
that
we
we
need
to
ensure
that
each
developer,
how
many
times
can
gay
can
go
online
so
so
for
each
year
we
can
do
about
twenty
thousand,
so
I
think
that's
I.
Think
one
of
the
key
words
is
that
each
person
can
each
person
per
day.
A
So
when
we
just
a
small
team,
it
was
easily
it
was
easy
to
you,
but
when
we
Scout
up,
then
the
distribution
volume
that
actually
we
need
to
maintain
that
volume.
It
was
quite
a
pressure.
It
was
quite
a
pressure
for
us,
so
Lin
trains
away
is
just
a
small,
quite
small
team,
and
now
we
can
have
a
direct
influence
on
our
team
members,
and
you
know
in
terms
of
DevOps.
It
wasn't
that
difficult.
A
We
our
work
these
project.
Actually,
the
goal
was
a
how
we
learned
their
DevOps,
so
we
also
have
away
from
our
clients
to
actually
be
one
of
our
clients.
Do
share
some
of
the
information,
so
I
will
briefly
go
through
the
background
of
the
project
and
then
also
the
overall
plan
and
construction
architecture
idea
of
our
project
and
the.
Lastly,
I'll
talk
about
the
project
and
technical
details,
so
for
companies
like.
A
A
So
when
we
talk
about
the
like
a
lot
of,
for
instance
or
your
fields,
we
have
dozens
of
all
your
fields
and
look
for
each
singer
of
oil
fields
day.
For
instance,
they
used
to
have
a
lot
of
systems.
Then
that
meant
that
we
had
like
several
hundreds
of
different
assistants.
So
it
was
difficult
to
manage
all
the
cysts
like
the
older
data,
and
so
we
couldn't
actually
be
able
to
have
a
great
support
for
the
transactions
so
this
time
that
we
use
these
platform
to
address
all
the
challenges
we've
faced.
A
So
actually,
this
is
just
a
picture
that
we.
So
this
is
just
a
brief,
like
architecture
of
our
project.
So
at
that
time
we
have
these
the
data
so
big,
so
today
we're
many
going
to
talk
about
DevOps.
So
we
have
these
paths
platform
and
we
have
three
scenarios,
including
container
platform,
DevOps
and
also
micro
service
governance.
So
the
three
important
core
scenarios
for
the
cloud
native
so
because
we
have
these
unified
platform,
so
we
have
some
support
for
the
business
of
patch
of
China's.
B
B
A
Time
scenarios,
so
these
so
as
mentioned
by
mr.
Chen.
So
it's
a
cloud
platform
based
on
container
divided
into
DevOps
micro
services
and
container
platform,
and
the
container
platform
is
kind
of
like
a
supporter
for
our
architecture,
and
we
have
some
tennis
system
in
that
have
safety
management
at
the
backend.
A
To
go
into
some
details,
we
have
eight
systems
for
our
DevOps,
so
for
like
we
have
some
tools
to
support
in
terms
of,
for
instance,
the
project
management,
because
for
each
project
we
actually
have
more
projects
than
products.
So
we,
when
we
have
a
project,
we
have
a
lot
of
regulations
and
rules
to
do
actually
tu
govern
all
the
projects
and
also
to
govern
their
business.
And
it's
kind
of
like
a
business
and
kind
of
like
an
experience
for
our
project.
B
B
A
Put
it
on
the
conference
and
folder
and
also
for
the
delivery
part,
so
it's
kind
of
like
a
code
and
we
have
deployments
you
run
all
the
activity.
So
it's
good
kind
of
like
an
automated
process
and
we
also
use
container
technology
to
support
our
automated
process
and
also
for
the
monitoring
that
the
DevOps
monitoring.
So
it's
not
only
about
the
project
we
have
develops,
so
it's
kind
of
like
a
monitoring
for
the
visualization.
A
You
know
so
for
this,
the
improvement,
understood
and
also
the
statistics
so
because
we
have
some
real-time
data
so,
for
instance,
the
requirements
that
so
how
like
how
many
success
we
had
and
the
failures
we
had.
So
we
had
all
those
data
and
have
these
kind
of
visualized
to
display,
in
that
you
have,
with
our
improvement
and
in
terms
of
the
material
I
write
a
materialization
and
she
provides
some
real
data
for
our
future
improvement.
A
And
you
have
this
plan
for
our
future
deployments
for
having
sustainable
and
sustained
and
in
the
future
and
a
for
the
training
site.
We
have
two
parts,
the
first
one
is.
We
have
a
training
and
learning
system,
and
the
second
part
is
that
we
have
actually
an
internal
environment
for
this
training
so
because,
since
last
year
we
started
these
modules
and
we
have
been
exploring
in
during
the
process
that
in
learning
and
improving
part
of
that,
training
takes
less
and
biggest
training
is
not
as
good
as
actually
doing
it.
A
So
we
still
think
that
practicing
is
better
than
just
lecturing
them,
because
we
have
a
certain
like
preset
code
and
the
pre
sad
environment
and
so,
for
instance,
within
two
days
of
training.
Maybe
it's
divided
into
one
and
a
half
day,
lecturer
and
then
half
day
or
maybe
one
day
for
practicing
and
we
were
have
an
trainer's
guide
them
to
actually
tell
them
how
to
put
those
theories
into
practice
and
we
have
a
DevOps
panaroma.
A
So
you
can
see
that
so
as
because
some
parts
I
have
already
covered,
because
at
the
bottom
we
have
a
accumulated
infrastructure
and
also
we
had
some
scenarios
which,
at
the
middle
level
that
we
can
go
into
some
details,
because
we
we
actually
have
these,
for
instance.
So
we
have
a
transaction
tracking
tool,
so
it's
not
just
about
a
single
tool.
A
B
A
Now
the
next
one
is
some
of
the
tools
that
we've
been
using
about
like
mood,
and
we
have
several
dozens
of
them
so
so,
like
we
have
like
aging
of
conference
and
next
for
the
coach
version
management
and
also
the
architecture
and
asuna
cube
for
the
static
code
and
sand
and
the
scanning
and
also
for
reading
rescue.
And
we
have
these
actually
testing
and
we.
A
One
and
asking
that
is
not
listed
here,
so
it's
also
for
testing,
so
for
that's
all
for
the
tours,
and
this
one
is
actually
the
procedure.
So
this
is
like
an
overall
procedure,
so
we
actually
have
a
lot
of
details
here.
So,
for
instance,
a
lot
of
like
we
have
requirements
and
demands,
then
we
have
a
detail
of
procedure
afford
these
and
it's
kind
of
like
procedure
that
we
have
some
summaries
and
we
actually
now.
A
We
have
some
changes
in
terms
of
the
procedure,
for
instance
for
a
normal
procedure,
we'll
have
five
different
Chandler
five
different
parts
and
when
we
put
it
into
actual
practice,
so,
for
instance,
if
you
kind
of
like
it
like,
if
it's
just
in
a
hurry
that
you
have
to
get
it
online
within
two
days,
so
like
five
steps
would
be
too
much
for
them.
It
takes
too
much
time
and
then.
A
A
So
since
we
adopted
to
the
micro
services,
we
only
have
a
few
people
working,
so
we
just
have
one
person
working
in
the
branch,
so
I
think
two
branches
will
be
enough
for
them,
and
now
this
is
the
rows
of
our
team
members.
This
is
only
part
of
that
and
because
we
have
for
the
DevOps
so
like
a
lot
of
yet
it's
just
about
the
development
and
also
users
and
also
practice
manager,
and
also
we
have
design
requirements.
A
A
Not
go
into
details
so
after
all
these,
so
we
will
have
orders
scenarios
so
like
what
rows
are
you
playing
and
what
tools
shall
we
use
and
what
regulations
and
what
procedures
shall
we
follow
that
you
realize
these
sooner
so
there's
kind
of
the
scenario
we
have
at
the
moment
and
all
the
full
parts
here.
So
it's
kind
of
like
a
more
detailed
scenarios
here
so
because
we
have
these
demand
or
management.
A
The
figures
like
the
four
scenarios-
and
this
is
kind
of
like
it's-
not
we
didn't-
make
it
up
figures.
They
are
actually
extracted
from
our
pain
points
that
we
encountered
like
when
we
ran
the
project
so,
for
instance,
the
developers-
maybe
they
felt
the
instruction
that
got
wasn't
clear
enough
and
then
also
like
the
dev
ops.
A
People
thought
that
maybe
they
didn't
have
enough,
like
mainly
the
data
was
like
to
schedule
around,
and
it
was
difficult
for
them
to
brown
the
projects
so
because
of
the
pain
points
that
we
have
actually
extracted
all
those
demands
and
realized
requirements
from
the
projects.
So
also
so
the
demand
management
scenario
we
have
for
in
different
scenarios,
so
all
the
way
from
the
mount
to
the
maintenance.
B
B
A
We
have
a
lot
of
like
detailed
regulations,
so
if
you
have
used
at
these,
so
you
know
that
one,
it's
quite
open
for
the
task
of
managing
and
we
didn't
have
like
certain
rules.
And
how
do
you
manage
it?
You
actually
have
like
flexibility
in
terms
of
the
running
of
the
project
and
in
terms
of
the
design,
and
we
have
the
project
manager
or
the
director.
So
when
they
finish
the
DES
past
and
we
have
these
developed
tester
scenarios
and
then
they
were
signed
to
the
developed
site
and
so
for
the
dev
site.
A
Then
the
developers
and
the
testers
they
were
gather
assignments
as
simultaneously
and
the
we
were
connected,
the
developments
and
the
months.
And
then
we
can
actually
see
that
the
project
that
we
have
like
what's
codes
that
we're
talking
at
and
what
projects
or
the
requirements
that
we
have
at
the
moment
and
also
I.
B
A
It
so
we
may
be
in
the
master,
they
need
to
submit
all
their
requests,
but
now
they
can
just
don't
realize
their
function
just
by
clicking
one
button
so
that
all
the
developers
and
the
testers
they
can
connect
each
other
and
they
can
share
the
data
and
also
after
when
we
finish
the
testing,
they
can
submitted
tasks
and
also
through
the
system
that
we
have.
This
may
be
all
the
offline,
so
we
have,
they
can
connect.
A
They
can
contact
the
test,
the
people
responsible
for
testing
and
then
those
people
will
follow
the
instructions
on
the
rows
and
inside
the
platform
there
was
extracted
scenarios
and
environment
and
because
we
have
different
like
dev
teams
and
we
have
a
certain
testing
environment,
so
it's
more
complicated
than
their
people
and
developers.
So
now.
B
A
A
simple
like
more
simple
way
for
running
this,
so
that
means
that
it's
like
more
like
easier
for
them
to
run
or
the
testing.
So
when
they
have
the
testing
environment
and
in
terms
of
the
like
the
functions
or
like
other
testings
after
finished
completed,
and
then
they
can
just
run
the
report.
And
after
what
afterwards,
they
can
just
send
a
report
to
the
people
working
in
the
DevOps
and
so
for
the
people
working
in
DevOps.
They
can
actually
like
the
operation
site
folder.
A
A
B
B
This
is
also
when
we're
doing
the
consultation
or
when
we're
doing
a
research,
we're
researching
what
can
this
container?
What
are
the
advantage
that
can
bring?
So
while
we
brought
here
today
are
the
real
case
scenario
and
what
kind
of
advantage
does
it
brought
to
us,
including
we
can
deliver
20
different
new
codes
per
day?
This
is,
we
cannot
imagine
this
in
the
past,
so
this
is
just
a
showcase
of
our
results.
B
There's
some
old
clay,
ok
scenarios
and
some
of
them
they're
real
and
they're-
also
other
statistical
with
inertial
forces.
How
many
para
lines
do
we
have
how
many
production
lines
do
we
have?
It
does
not
show
over
here-
and
this
does
not
look
so
big
today,
and
here
is
also
our
example
and
results
for
our
procedures.
B
B
Thank
you.
So
if
I
want
to
talk
about
example,
it's
better
leave
that
to
our
customers
that
topic.
Actually,
he
can
talked
about
half
an
hour's
or
even
more,
but
to
the
time
limits
and
we're
gonna
talk
too
much
about
that.
So
you
can
feel
like
for
a
traditional
enterprise.
They,
like
him,
China
petrol.
B
B
B
Ac
is
basically
for
these
huge
entities.
Customers.
As
for
the
China
petrol
we've
mentioned,
the
rate
doll
is
is
actually
very
suitable
for
this
platform
is,
he
includes
all
the
characteristics
of
a
CP
that
is
for
establishing
a
unilateral
platform
for
huge
enterprises,
because
some
of
these
big
customers,
they
have
very
complicated
infrastructures
environment.
B
So
when
people
when
people
starting
to
use
public
health,
they
will
not
focus
on
one
cloud
provider,
they
will
use
multiple
cloud
providers,
so
we
have
offered
this
multi
cluster
management
and
multi
cloud
management
function.
So
we
can
supply
this
multiple
cables
cluster
and
we
can
manage
very
complicated
cluster.
This
is
what
we
can
do
beyond
a
managing
of
the
cluster
can
not
only
through
AC
e.
B
B
B
From
business
unit,
perspectives
is
quite
complicated
for
the
AC
use
adoption
scenarios.
We
have
thousands
of
different
scenarios
and
between
different
that
scenarios
we
need
to
have
different
types
of
renting
customers
and
they
need
to
work
collaboratively
and
sometimes
they
need
to
have
this
isolation
functions
for
the
resource,
and
these
are
customers
asking
for
it,
and
this
is
what
a
CEO
needs
to
solve.
So
today
we
need
to
talk
about
mainly
we're
gonna
talk
about
DevOps.
B
This
is
the
main
module
in
a
CEO,
so
we're
when
we
are
talking
about
these
DevOps
requirements
with
our
customers
they're
already
using
some
tools
inside,
and
there
are
some
procedures
that
we
need
to
complete
or
even
to
finalize
some
product
that
we
already
exist.
We
cannot
just
simply
replace
them
because
it's
not
suitable,
so
this
is
actually
the
thinking
map
that
we're
using
right
now.
So
this
is
a
integration
that
we
can
call
through
integration
and
cluster.
B
We
can
link
everything
together
into
a
to
link
and
to
become
as
a
whole,
so
we
can
manage
it's
through
its
through
life
range
and
through
automation.
We
can
serve
the
platform
and
to
offer
our
customers
the
best
service.
This
intake
is
integration
of
DevOps.
We
can
deploy
this
on
our
container
platform.
This
is
not
necessary.
Sometimes
it's
not
necessary
because
some
container
is
already
exist.
If
we
have
this
public
sauce
service,
we
can
just
put
it
into
the
public
cloud
platform.
B
But
one
thing
we
need
to
make
sure
is
that
we
need
to
have
the
platform
and
tooltip
of
integration,
and
also
we
need
to
the
platform
into
a
premises.
We
need
to
link
that
together,
the
SSO
each
enterprise
tool
will
have
its
model
for
our
inter
customers,
and
also
these
needs
to
have
linkage
between
each
other.
For
instance,
we
need
to
thousands
of
production
engineers
in
it
and
to
share
these
tools
together,
and
they
need
to
follow
some
certain
regulations
and
finally,.
B
B
B
B
The
infra
is
for
them
is
prepared
by
the
manager
of
the
platform,
and
then
they
will
create
these
rental
managers
on
the
platform.
They
will
allocate
the
resources
to
each
users
relocate
them,
what
which,
what
amount
of
resource
they
can
use,
and
this
user
management
after
they
have
acquired
this
resource,
they
will
create
their
own
environment.
That
is
the
name
space
within
the
Cuba
natives,
but
some
users-
they
are
so
huge.
Well,
they
can
be
tens
of
thousands
of
them,
so
it
can
be.
B
They
can
be
quite
big
and
under
each
users
we
can
create
different
user
groups
and
this
environmental
resource.
They
can
also
allocate
to
these
small
groups
under
the
users
as
well
for
DevOps.
They
can
have
integration
and
cluster
of
resource.
They
can
use
some
tools,
these
tools,
you
can
see
that
as
a
service
that
can
be
published
and
the
users
that
can
subscribe
them
as
for
China
petrol.
This
is
the
modes
that
we
use
and
they
can
establish
this
resource
for
different
tools
and
allocate
them
to
different
users
from
the
users
manager.
B
There
are
three
ways
they
can
acquire
resource.
Maybe
they
can
directly
get
it
from
the
resource
manager.
Maybe
they
can
subscribe
from
the
platform
manager
also,
they
can
deploy
the
tool
by
themselves,
so
they
can
establish
their
own
resource
in
their
own
user
group.
They
can
also
allocate
their
resource
to
the
small
groups.
That's
underneath
them,
and
these
are
the
details.
On
the
left
hand,
side
you
can
see
that
they
are
actually
allocating
some
cluster
resource
to
our
users.
There's
a
cluster
called
the
Federation
one.
B
B
B
So
when
our
users
deploying
this
Federation
here,
you
will
ultimately
deployed
into
two
different
clusters
later
on.
They
can
manage
this
independently,
but
overall
they
can
manage
this
together.
This
is
specially
useful
within
the
financial
industry.
On
the
right
hand,
side
is,
after
choosing
the
cluster.
You
can
allocate
the
resource
within
the
cluster
for
our
users,
it's
basically
very
similar
to
the
name,
the
name
cell
and
the
namespace
function,
this
user
manager.
B
They
can
establish
groups
under
the
users
if
this
user
group
is
quite
big
but
more
importantly,
they
need
to
establish
their
own
environment
and
that
environment
is
the
namespace
under
quba
natives.
We
can
also
configure
the
resource
that
they
need
that
may
want
to
use.
We
know
that,
under
this
big
cumulative
system,
we
can
we
have
layers
and
layers
of
different
resources
that
we
can
allocate,
but
it
cannot
surpass
the
resources
overall
allocated
by
the
upper
layer.
B
B
B
After
establishing
this
resource,
they
can
allocate
these
to
the
users
and
these
users
they
can
check
out
what
type
of
resource
has
been
published
by
the
manager
and
they
can
subscribe
it
and
they
can
use
also
they
can
create
their
own
resource
as
well
and
to
allocate
them
into
their
small
group.
So,
for
the
final
five
minutes
I
would
like
to
introduce.
How
do
we
do
CI
CD
through
Jenkins
and
kubera
natives,
so
we
have
chills
in
some
very
deep
CI
CD
examples.
There
are
two
main
goals
when
we're
designing
this
platform.
B
The
first
goal
is
that
maybe
our
customer,
they
have
the
small
group
of
people
within
their
enterprise.
I
know
about
Jenkins,
but
most
of
the
developers
for
CI
CD.
They
do
not
know
a
lot
about
Jing
as
file
if
it's
too
complicated,
if
there's
no
one,
if
no
one's
knows
about
Jenkins
that
we
can
also
offer
our
help
for
design
and
the
other
target
of
our
when
designing
is
that
we
need
we.
B
Hopefully
we
can
cover
80%
of
our
design
scenarios
on
the
platform,
so
there
was
only
a
small
chance
that
need
to
go
into
tools
and
check
out
their
own
tools.
So
basically
they
can
finish
this
by
themselves
for
production
lines.
They
can
establish
their
own
production
line
and
they
can
have
this
mutual.
They
can
send
out
this
mutual
commanding
and
check
out
their
logs.
They
can
check
out
the
results
and
of
they
all
can
take
a
note
of
debug,
so
we
can
have.
B
B
This
Cuba
natives,
we
we
have
used
the
Jenkins
plus
crew,
very
natives
and
the
alota
and
Jenkins
plug-in
or
the
DSL
to
manage
it,
and
we
support
this
create
this
creation
of
production
line.
We
know
that
most
of
our
users,
they
will
not
touch
and
can
file
every
time
when
they
want
to
create
this
production
line.
B
B
B
B
B
So
under
Jenkins
we
have
this
plugin
for
ultra
sign
up
and
also
we
can
have
this
auto
synchronize
authentication
plugin.
We
know
that
under
Jenkins
it's
a
metrics
and
on
our
platform
is
based
on
the
user's
role,
so
depend
on
that
role.
We
can
offer
its
authentication
and
privilege,
and
it's
all
under
the
Kuban
native
system.
We
use
that
as
a
framework,
including
like
the
production
line
in
the
pipeline.
B
If
it's
for
API,
we
just
offer
that
through
the
Cuba
native
server
and
other
tools,
we'll
have
customer
controller,
and
finally,
is
that,
hopefully
we
can
link
all
the
stage
different
stages
within
our
products
lifetime,
which
will
be
easier
for
them
to
deploy,
for
instance,
this
production
line
they've
run
this
test.
We've
finished
that,
but
after
the
submit
of
our
coding
is
for
this
times
development
mission
and
this
mission
is
based
on
the
requirements
of
the
development
of
this
customers
development.