►
From YouTube: Community Stream #22: Anthony Miloslavsky
Description
Anthony Miloslavsky talks Automated Network Testing. Follow Anthony @permitanyany on twitter, and check out his CHINOG presentation on automation testing here: https://www.youtube.com/watch?v=T_pUhLsAYOY
A
Oh
hello,
everyone
I,
am
cloud
toad
at
Cielo,
UDT,
Oh
ad
on
Twitter
and
I.
Am
your
host
today
on
the
NRI
labs.
Livestream
enemy
labs
is
a
free
and
open
source
project
which,
with
the
the
it's
only
purpose,
is
to
teach
people
about
Network,
automation,
automation,
concepts
for
network
engineers,
etc,
etc.
You
can
follow
that
project,
add
enemy
labs
on
Twitter
and
you
can
go
to
the
website.
Labs
network,
reliability
at
engineering,
and
you
can
check
out
the
content
for
free,
there's,
no
marketing
wall,
you,
don't
you
don't
have
to
identify
yourself
for
anything.
A
A
We
practiced
that
before
we
started
and
I
still
managed
to
screw
it
on
he's
done.
He
not
that
long
ago,
in
Chicago,
at
the
shine
nog
did
a
really
interesting
talk
on
the
importance
of
testing
in
network
automation
and
the
danger
of
not
doing
tests
and
testing
there
something
I'm
also
very
familiar
with.
When
you
don't
do
testing
with
network
automation,
you
can
really
hurt
your
network
badly.
It's
like
the
blast
radius
becomes
much
bigger
and
much
faster
without
without
testing
your
automation,
the
automation
things
that
you
build.
B
Sure,
thanks
so
I
started
networking
actually
a
while
ago
and
my
junior
year
of
high
school,
my
high
school
happens
to
be
giving
out
the
Cisco
Academy
courses
as
part
of
a
required
curriculum,
so
that
was
kind
of
my
first
foray
into
networking
concepts
to
be
fully
honest.
I
didn't
quite
follow
what
they
were
teaching
I
didn't
really
understand
the
big
picture,
I
kind
of
learned,
isolated
facts,
you
know
so
on
and
so
forth,
which
only
many
years
later
I
would
kind
of
together
and
fully
understand.
A
B
It
kind
of
started
following
you
know,
following
the
trends
starting
to
do
some
of
that
on
my
own
or
not
they're,
trying
to
script
day-to-day
tasks,
to
make
my
job
a
little,
less
boring
and,
at
the
same
time,
I
got
involved
in
our
public
cloud
project
or
we're
pulling
infrastructure
into
AWS
and
honestly,
when
I,
when
I
saw
what
type
of
automation
and
how
that
team
was
using
infrastructure
as
code
to
deploy
various
aspects
into
AWS.
That
kind
of
just
doubled
down.
B
B
Can
we
do
some?
We
add
some
automated
testing
into
this,
even
if
it's
basic
and
the
response
I
got
was
just
identical
that
it
seemed
throughout
the
industry.
You
know
this
is
all
so
new
to
me.
Why
don't
we
focus
on
the
configuration
part
first
and
we'll
do
the
automated
testing
part
after
that,
and
that's
just
really
been
the
pattern
that
I've
seen
throughout
through
either
our
industry
and
I
think
I
think
talking
about
automated
testing
how
to
do
it?
A
Absolutely
you
know
we
actually
one
of
the
things
I've
said
on
this
kind
of,
certainly
on
this
podcast
before
and
almost
in
every
presentation.
I
give
is
that
you
know
we
spend
most
of
our
time
actually
modeling
as
network
engineers
like
we're,
trying
to
build
a
picture
of
what
is
happening
or
what
should
be
happening
or
what
did
happen
right,
remodeling,
the
past
the
present
or
the
future,
so
that
we
so
that
we
make
the
correct
choices
when
we
actually
go
and
try
to
manipulate
the
system
right.
A
If
we
don't
have
a
good
working
model
in
our
head
of
where
we
are
and
where
we
should
be,
then
then
things
can
go
very
badly
right.
That's
that's
the
definition
of
incompetence.
We
spend
a
lot
of
time
actually
doing
a
lot
of
modeling,
so
it
seems
if
you're
only
focusing
on
configuration
you're.
It's
that's
actually
kind
of
a
scary
thought
right.
The
danger
that
can
go
with
that
I've
taken
out
an
entire
network
way
back
early
in
my
career.
A
Why
just
focusing
only
on
that
and
took
out
94
branches
of
a
bank,
and
it
took
three
full
days
of
trucks
rolling
to
to
undo
what
I
did
because
I
just
I,
you
know
I.
Now
it
wasn't
entirely
my
fault,
but
I
should
have
known
better
before
you
know.
I
did
that
that
each
you
know,
even
if
you're
doing
a
lot
of
people,
don't
get
it.
How
this
is
how
important
it
is
right.
A
Even
let's
say
you
have
a
thousand
devices
in
your
network
and
they're
all
the
exact
same
model,
the
exact
same
memory
running
the
exact
same
iOS
every
single
time.
You
should
be
validating
the
change
you
just
made
on
that
box
and
then
it
make
sure
you're
running
the
right
version
of
iOS
right
that
it's.
That
is
in
fact
the
same
thing
that
you
tested
in
the
lab
before
you
make
the
change
and
then
validate
it
after
you
make
the
change
for
each
and
every
of
those
1,000
devices.
B
Yeah
agreed
in
I
mean
I
could
think
back
to
numerous
times
of
my
previous.
You
know
enterprise
jobs.
Where
you
know
we
we
had
to
do
late
night,
late
night,
McMahons
windows,
no,
usually
the
more
impactful,
the
more
off-peak
it
has
to
be.
You
know,
sometimes
in
the
middle
of
the
night,
on
a
Saturday
and
at
those
times
you
know
if
you're,
if
you're
a
couple
hours
into
troubleshooting,
something
we're
deploying
something
you.
B
Your
head
is
not
fresh
in
terms
of
you
know.
Let
me
let
me
check
out
all
the
all
the
changes
I've
made.
Let
me
do
some
due
diligence
and
you
know
further
into
that
and
maybe
you're
four
hours
into
it.
You
know
your
brain
is
just
I
want
to
get
out
of
here
as
quickly
as
possible
right,
so
you
know
focusing.
A
B
A
B
You
know
humans,
they
can
only
go
for
so
long.
You
know
in
terms
of
looking
at
things
you
know,
especially
if
it's
happening
it's
happening
manual,
so
I
think
I.
Think
there's
there's
so
many
different
avenues
to
look
at
automated
testing
and
I
really
think
it's
there.
No
brainer
in
each
section,
I
think.
A
B
B
Are
at
night
you
know
we're
human.
After
all,
the
leader,
you
are
into
a
meetings
window,
the
more
high-stress.
It
is
the
more
you
want
to
get
out
of
there
kind
of
get
back
to
go
back
to
normal
weekend
life.
That's
really
where
automated
testing
can
help
right.
Just
rely
on
that
suitable
tasks,
help
you
to
make
sure
make
sure
you
change.
You
went
fine
and
I.
You
know
I
could
think
of
many
times
where
I
did
all
that
manually
and
obviously
many
times
miss
something
just
because
it
was
all.
A
Yeah,
you
know,
like
you
said,
when
you're
at
your
peak
mental
faculty,
yet
you
know
3:00
in
the
morning
on
Sunday.
You
know
when
you're
you're
stressed
out,
because
it's
not
like
you
know,
you're
missing
a
big
chunk
of
your
personal
life.
At
the
same
time,
by
the
way
you
know
you're,
perhaps
foggy
or
yeah
peak
mental
faculties,
so
automated
testing
is
good
because
you
know
it's
some.
These
things
are
complicated
right
now.
Sometimes
it's
not
just
what
you're
doing
on
this
box.
A
It's
like
the
effect
of
the
box,
that's
connected
to
that
or
a
box.
That's
even
downstream
from
that
right,
because
it's
a
network,
it's
like
one!
It's
actually
one
sort
of
large
distributed
system.
That's
isn't
that
crazy
that
we
still
do
that.
Like
there's
like
one
of
the
most
complicated
things,
our
infrastructure,
we
decide
to
do
all
the
moat,
the
most
difficult
stuff
when
we
have
the
least
amount
of
brain
power,
no.
B
Yeah,
it's
ironic
I,
remember
even
I,
remember
you
know
and
change
control
with
these
large
enterprises.
You
know
one
of
the
common
questions.
As
always,
how
are
you
going
to
verify
that
your
change
was
successful
right
and
that
kind
of
leads
us
into
a
rabbit
hole
with
its
own
conversation,
but
yeah.
B
A
So
so,
what's
sort
of
so
you
said
you've,
you
know
you've
sort
of
gotten
into
scripting
or
you're
doing
well.
We
call
it
scripting
rain
network
engineers,
frequently
call
it
scripting,
but
it's
it's.
You
know
it's
definitely
evolved.
It's
not
just
scripting
anymore.
There's.
You
know,
there's
a
practice
around
it
so
like
what
what
sort
of
languages
and
tools
were
using
in
the
beginning,
yeah.
B
So
I
think
my
very
first
forays
into
this
was
and
kisses.
No.
This
is
that
key
back
was
a
expect
scripting
right,
it's
kind
of
the
best,
that's
what
we
had
back
then
and
I
remember
I.
Think
one
of
my
one
of
my
first
managers
actually
asked
me
to
you
asked
me
to
build
the
equivalent
of
zero
touch
provisioning.
You
know
with
whatever
it's
what
we
had
back,
then
that
way
we
can
easily
provision
switches
at
least
get
them
on
the
network
get
them.
B
B
You
know
it
really
depends.
I
think
the
testing
piece
is
still.
You
know
it.
There's
no
jury
still
Adam.
What
what's
the
easiest
way
to
do
it,
but
I
think
I.
Think
for
starters,
it's
it's
nice
to
run
a
couple
of
commands.
I
think
where
it
gets
complicated
is
in
my
personal
experience,
it's
a
lot
easier
to
parse
output
using
just
native
Python,
at
least
it
is
to
me
and
if
I
know,
I'm
going
to
be
doing
a
lot
of
parsing.
A
B
B
Application
team,
an
abstract
and
test
that
you
doing
on
the
network
side
without
without
them,
really
having
to
think
about
what's
happening
behind
the
scene
and
vice
versa.
Right,
if
you
want
to
make
a
network
change
and
you
want
to
get
an
abstracted
application,
you
know
test
check
out
from
them
without
really
having
to
think
about
what's
happening
behind
the
scenes.
Robot
seems
really
interesting
in
that
area,
so
I
played.
B
So
there
was
mostly
mostly
Python
and
ansible,
and
you
know
the
main.
The
main
difference
I
kind
of
saw
there
is
that
anthem
will
definitely
definitely
attract
abstract,
a
lot
of
complexity
for
you,
we
at
least
on
the
surface,
but
the
great
part
about
that
is.
If
you
have
a
large
team
and
you
have
varying
degrees
of
expertise
with
automation
and
some
was
pretty
pretty
easy
to
collaborate
on
and
relatively
easy
to
read
right
I
can
I
can
write
something
in
ansible
and
chances
are.
B
I
can
explain
to
my
coworker
now
doing
that
with
a
Python
script,
that
I
wrote,
that's
that's
much
harder
and
I
mean
I.
Think
I
think
you
can
pop.
You
can
probably
agree
that
a
lot
of
those
scripts
go
to
die
when
that
person
leaves
the
company
right
so
or
at
least
you
know,
everybody
still
uses
them
until
they
break
and
when
they
break
nobody
really
knows
how
to
fix
them.
So
I
think
that's
one
thing
to
like
anthem
will
have
going
for
them,
but
yeah.
A
Yeah
I
mean
so,
is
this
testing
the
the
sort
of
automation
testing
you
do?
Is
it?
Is
this
something
that's
built
into
your
organization's
process,
or
is
this
or
your
sort
of
the
guys
sort
of
spearheading
this
and
trying
to
you
know,
get
it
more
widely
adopted?
We
run
into
a
lot
of
the
latter,
rather
than
the
former
yeah.
B
B
Like
I
said,
I'm
gonna
kind
of
I'm
on
the
vendor
side
today
and
so
now,
I'm
talking
to
clients
that
are
adopting
this
or
I'm
helping
them
adopt
this,
and
it's
still,
you
know
it's
still
a
mixed
bag.
Someone
gets
totally
behind
it,
they
want
and
they're
they're
great
are
contributing
to
it
and
others.
Just
you
know,
they'll
take
whatever
code,
I
give
them
and
that's
kind
of
gonna
be
the
stock.
You
know
the
stock
testing
that
they'll
run
and
that's
it
so
yeah
both.
A
Will
make
spec
for
sure?
Do
you
have
any
advice
for
someone
who's
trying
to
get
a
practice
going,
especially
in
an
organization
where
it's
you
know,
they're
they're,
largely
surrounded
by
people
who
are
just
you
know
this
is
completely
foreign
and
that
they
resist
rain.
Do
you
have
any
advice
for
them?
Yeah.
B
B
It's
really
it's
pretty
much
risk-free,
so
you
know
what
I
did
when
I
started
this
in
my
previous
organization,
that's
kind
of
exactly
what
I
did
you
know.
I
I
took
these
large
complex
changes
or
changes
that
people
make
more
than
once
and
just
wrote
something
in
Python.
You
know
I'm
sure
it
was
mostly
readable
to
me,
but
so
and
just
get
something
repeatable
that
you
can
use
so
I
think
it's
a
really
risk-free
easy
way
to
start
and
I.
A
A
So,
what's
sort
of,
let's
say
someone's
you
know
watching
this
and
they're
like
I
wanted
you
know.
Testing
sounds
like
a
great
way
to
get
is
I
by
the
way.
I
say
that
all
the
time
I
tell
everyone
your
first
automation
should
not
be
doing
configuration,
it
should
be
pulling
information
and
munging
it
and
then
presenting
it
in
a
consumable
way
and
I.
One
example
I
like
to
give
there's
a
lot
of
places
that
have
like
vrrp
or
HS
RP
or
whatever
in
their
environment,
and
you
know
that's
one.
A
That's,
oh,
that's
a
great
first
thing
that
you
can
sort
of
point
a
script
at
download
all
information
in
one
block
like
here's,
all
the
participants
in
this
verb
domain
or
whatever
and
I
you
can
see
all
their
status
and
all
the
times
I
that
their
status
has
changed
whatever,
and
it's
a
it's
like
a
great
first
project.
Any
suggestions
for
preferred
for
that
kind
of,
like
let's
say
modeling,
someone
could
do
like
first.
First
steps.
B
B
As
it
should
be
right,
you
have
your
you
know:
you
have
their
preset
NTP
servers
or
SNMP
servers
on
all
of
your
devices,
so
you
know
even
starting
with
something
as
simple
as
that.
You
know
if
you
have
a
thousand
thousand
switches,
just
logging
into
every
one
of
them,
pulling
their
SNMP
n
ntp
status
and
making
sure
it
is
what
you
expect
it
to
be,
and
so
you
know
taking
that
building
upon
it.
To
kind
of
you
know,
make
post
change
checkouts
right.
A
A
A
B
B
A
I
agree
so
when
you're
can't
stress
this
enough
and
I'm
sure
you
can
why
don't
you
speak
to
that
I
think
the
importance
of
like
these
especially
user
kind
of
like
shy,
nog
and
I
nog
or
really
about
operators.
That's
right
in
the
name,
ray
network
operators.
Group
I
would
argue
that
you
know
community
online
communities.
Great
right,
I
mean
it's.
You
know
when
software
isn't
destroying
your
your
live
stream.
A
You
know
it's
all
the
stuff
that
can
happen
online
forums
and
podcasts
and
live
streams
and
Twitter
and
and
all
it
you
know
it's
all
kinds
of
slacks
and
IRC
channels
and
everything
now
discord,
servers,
etc.
But
you
know
going
to
these
events
and
meeting
people
in
person.
I
think
is
really
important.
You
talk
about
that.
A
little
bit
yeah.
B
And
I
mean
you
know
we
have
you
know
we
have
access
to
this.
You
know
Network
community.
However,
however
big
it
is
whether
it's
on
social
media
or
on
flack,
but
you
know,
there's
a
lot
of
people
that
a
maybe
aren't
on
that
community
and
they
just
and
their
local
events.
And
then
you
have
the
opportunity
to
run.
B
B
You
know,
I
I,
think
it's
I,
think
it's
great
for
your
career
I.
Think
it's
great
to
get
kind
of
a
perspective
of
you
know
what
are
other
engineers
in
this
field.
Doing
so
I
think
you
know
just
professionally
to
kind
of
grow
and
see
what's
out
there.
It
gives
you
a
nice
nice
view
into
to
what's
going
on
so
I
I
personally.
A
And
you
know
the
great
thing
about
that
is
there's
so
much
you
can
find
it
online.
But
absolutely
when
you
go
in
person
you
get
to
there's
something
about.
You
know
you,
it's
a
it's
a
lot
more
rewarding
when
you
know
you
sit
and
you
listen
to
someone's
talk
and
you
can
ask
questions
or
you
know,
catch
up
with
the
person
afterwards
and
ask
questions,
and
you
know
you
get
a
lot
of
great
pointers
and
tips.
B
They're,
you
might,
you
might
think
of
them
as
celebrities,
but
they're
really
not
we're
all
in
this
kind
of
small
small
field
together
and
they're
very
accessible.
They
will
always
answer
your
questions.
That's
just
in
my
experience,
so
it
could
definitely
be
intimidating
approaching
somebody
or
even
asking
a
question
in
public
but
I
think
for
the
most
part
you
know
it's.
It's.
A
B
A
A
B
A
A
You
know
we're
we're
improving
the
tooling
and
stuff
for
actually
creating
lessons
on
the
site,
we're
getting
other
we're
pulling
all
the
vendors,
and
so
you
know
cumulus
gave
us
an
appliance.
We
have
an
fr,
our
appliance
now.
The
free
range
routing
appliance
cumulus,
is
based
on
fr.
Our
but
running
a
Linux
machine
with
FRR
and
running
cumulus
are
two
completely
different
things
correct.
So
we
we
have
both.
So
we
we
do,
have
you
know
multiple
vendors
and
then,
of
course,
we
have
vendors
outside
a
networking
ray.
We
just
were
working
with
Red
Hat.
A
Now
we're
gonna
have
ansible
tower
up
here
soon
in
a
you
know,
and
other
tooling,
in
etc.
So
I
I
think
it's
it's
it's
going.
The
direction
we
wanted,
I,
think
being
multi
vendor
and
being
open
and
free
without
a
marketing
wall
has
been.
Is
the
right
way
to
do
it
and
I
think
it'll?
You
know
the
more
we
you
know
become
more
like
and
an
open
source,
an
official
sort
of
open
source
community.
We
are
open
source
totally,
but
you
know
we're
moving
into,
possibly
the
software
free
Conservancy
will.
A
Hopefully
you
know
we'll
see
how
that
unfolds,
but
they're
gonna
evaluate
that
pretty
soon,
and
you
know
we
got
a
our
governing
stock
in
our
code
of
conduct
and
we
have
all
these
other
things
and
I
think
is
we.
You
know,
as
all
that
gets
settled
and
as
we
get
other
vendors
to
participate
and
I
think
this
project
will
start
to
grow
even
faster,
and
certainly
when
we,
you
know
we
get
more
of
the
tooling
straightened
out
for
for
creating
the
lessons
you
can
create
a
lesson.
A
Anyone
can
create
a
lesson
now,
but
it
does
take
a
little
bit
of
linux
knowledge
to
get
the
development
environment
set
up
and
to
make
your
own
images
for
the
for
the
lessons
and
stuff
like
that,
that'll
we're
you
know,
that's
that's
a
key
focus
first.
This
next
year
is,
is
making
those
tools
even
easier
to
use
so
that
you
don't
you
know,
spend
a
lot
of
time
crane.
A
Those
images
that
for
our
image,
I
created
myself
and
that
took
some
time
some
time
to
actually
get
functional
and
it
needs
to
be
better
than
that,
but
I
think
we're
I.
Think
we're
at
where
we
need
to
be
people
are
asking
the
right
questions.
People
are
using
the
lessons
anywhere.
We
launch
anywhere
from
400
to
800
lessons
a
week.
Yeah
I
think
it's
coming
along
good
I.
Think
people
are
interested.
People
wanted,
people
want
to
know
more
and
know.
A
Hopefully
this
your
I
kind
of
want
to
make
sure
some
of
the
lessons
are
more
focused
on
testing
and
modeling
and
less
on.
You
know,
but
we're
gonna,
you
know,
push
out
changes,
because
that's
that's
what
I
mean
that's
what
we
do
most
of
our
time.
We
should
make
it
easier
for
us
to
model
things,
especially
at
2:00
in
the
morning.
B
Yeah
yeah
well,
one
one
piece
I
was
gonna
address
is
that
you
know
I
personally.
The
way
I
learned
is
by
doing
and
I
think
I
think
that
applies
flights
and
most
engineers.
So
what
I?
You
know
when
I
hear
the
question
or
what
I'm
talking
you
know
it's
a
kind
of
Engineers
that
aren't
as
familiar
with
automation.
The
topic
of
how
do
I
get
started
always
comes
up,
and
you
know
if,
when
my
answer
is
what
you
have
to
learn
by
doing
you
have
to
do
this,
you
have
to
do
that.
B
I
feel
like
it.
You
know
not
everybody's.
We
do,
or
at
least
they
aren't
until
they
start,
and
they
see
the
benefit
of
it.
So
you
know
your
your
platform.
Energy
labs
definitely
helps
in
that
case,
where
I
could
just
you
know,
point
them
to
it
and
say:
hey
go
to
this
site.
Take
this
lesson
right,
and
at
least
you
know
at
least
you're
getting
started.
You're
you're,
lowering.
A
A
Yeah,
that
was
the
goal.
That
was
the
goal
it
is
learning
by
doing
is
absolutely
learning.
By
doing
it,
a
lot
of
network
engineers
are
sort
of
autodidactic
anyway,
in
a
way
I
said,
I
brought
this
up
last
week,
like
there's,
there's
not
really
I
mean
there
is
network,
engineer,
training
that
you
can
buy
from
vendors
and
from
third
parties,
but
you're
not
going
to
be
a
bgp
expert.
A
Even
if
you
like,
there's
advanced
bgp
training,
you
can
take
from
multiple
vendors
until
you
do
it
a
lot
right
until
you
really
see
you
know
hey,
why
are
we
adding
extended
communities
to
this
bgp
route
over
here
and
then
we're
matching
on
it
over
here
and
then
manipulating
it
like
you
until
you've
done
that
in
production
and
you
sort-
and
you
start
to
get
a
feel
for
that
kind
of
thing-
you're.
Not,
then
you
realize
there's
so
much
about
BGP.
That's!
Never
in
any
of
the
training.
Right
and
automation
is
even
bigger.
A
It's
even
bigger
I
mean
there's,
there's
so
many
tools
and
languages,
and
it's
it's
it's
bong
and
syntaxes
right,
and
it's
you
it's
it's.
Even
if
it's
the
universe
is
even
larger
and
you
can
take
some
automation,
training,
but
it's
you
know
it
you're,
you
just
you
have
to
do
it.
You
have
to
do
it
or
you're,
never
really
going
to
be
you're.
Never
gonna
get
good
at
it.
B
A
B
A
The
whole
time,
but
there
we
go
hey
it's
fixed
now,
Wow.
Well,
there's
people
watching
and
they're
commenting
I
just
want
to
say:
I,
we've
had
a
cursed
stream.
So
far
the
we
were
using
zoom
for
that
for
the
client
and
it
restarted
in
the
middle
of
this
middle
of
the
conference
because
it
updated
itself
and
then
and
then
my
headset
died
and
then
yeah
and
now
his
audio
stopped.
With
your
your
back
now,
they're
saying
they
can
hear
you
now:
okay,.
A
Weird
I
would
say
that
there's
two
things:
it
starts
off
with
one
or
two
or
three
people.
Let's
say
who
get
a
who
get
a
this
idea
that
you
know
hey?
They
have
the
vision
right,
they'd
like
it
strikes
them
that
automation.
We
ought
to
be
doing
it
because
we're
doing
all
these
changes
at
2:00
in
the
morning,
there's
like
especially
you
know,
I
would
say
especially
a
medium
or
large
shops
is
where
they
really
see
it
because
tangibly
they
feel
they
feel
the
side
effect
of
not
doing
it.
A
One
is
when
they
realize
that
testing
is
more
important
than
configuration
changes,
which
is
why
I
just
tell
people
right
off
the
bat.
Don't
worry
about
automating
your
configuration
changes
if
you're
not
becoming
really
good
at
modeling.
What's
going
on
in
the
network?
First
then
you're
setting
yourself
up
for
catastrophic
failure.
There's
I
mean
you're
gonna,
push
out
a
change
to
a
thousand
devices
and
you're
gonna
blow
up
your
network,
the
guy
I.
They
just
start
by
modeling.
A
It's
what
you
spend
all
your
time
on
anyways,
so
when
they
have
that
realization
that
testing
and
modeling
is
is
really
the
most
important
thing
to
automate
because
look
the
information
is
complicated.
Networks
are
complicated
like
how
can
you
pop,
even
when
you
have
all
the
certifications
II's,
you
know,
I
had
three
II's
at
one
point
like
you
can't
all
that
in
your
head,
especially
you
know,
especially
at
2:00
in
the
morning,
so
testing
and
modeling.
They
have
that
realization.
A
You
fix
the
process
right
and,
and
you
stay
on
top
of
that
and
that's
DevOps
right.
It's
the
process
of
the
process
when
so
when,
when
they
realize
testing
and
modeling
is
important
and
when
they,
when
management
gets
involved,
and
they
start
to
formalize
it
and
make
it
part
of
what
they
do.
That's
that's
when
it
really
takes
off
until
then
it
you
know
it's
challenging
because
you'll
end
up
with
someone
I've
been
in
this
position.
You've,
just
you
said
you
were
in
that
position
earlier.
A
You
know
you're,
you
know
you're,
it's
like
banging
your
head
on
a
wall
like
you're
you're,
trying
you're
trying
to
make
it
happen
and
you're
showing
people
the
benefits
and
and
look
you
know,
here's
the
results.
You
know
I
ran
this
change,
I
did
it.
It
was
good
on
every
device
and
I
also
looked
over
here
to
make
sure
those
routes
appeared
or
whatever
and
I
and
I
ran
this
report,
and
here
it
is
like
it
just
happened
and
I
didn't
need
ten
guys
pounding
away
to
ten
people
pounding
away
on
a
keyboard.
A
A
So
you
ought
to
be
selling
this
upward
as
you're
doing
it
as
much
as
possible
and
the
pain
point
for
management.
If
I'm
gonna
give
any
advice
and
keep
talking
while
you're
the
guest,
is
you
know,
managers
don't
like
this
is
a
lot
of
network
engineers
love
how
frustrating
it
is
being
a
network
engineer,
sometimes
because
nobody
really
understands
what
you
do
is
stuff
for
other
network
engineers,
but
I
want
to
just
put
yourself
in
the
position
of
somebody
who
manages
a
network
team
when
you
blow
something
up
in
the
network.
A
They
first
of
all
may
not
really
understand
what
you
did.
That
was
wrong
right
there.
Their
job
is
to
manage
the
process,
not
to
be
a
network
engineer
and,
and
they
have
they
have
to
answer
to
customers
and
into
up
the
chain
in
your
organization
and
sometimes
when
you're
trying
to
explain
to
them
something
really
arcane
about
what
you
did
that
cause
that
outage
and
they
have
to
turn
around
and
explain
to
people
what
the
root
cause
of
the
problem
was
and
that
it's
never
gonna
happen
again,
which
is
a
lie.
A
A
You
walk
into
a
you
know,
walked
into
the
office
of
a
guy
whose
job
is
to
get
yelled
at
and
for
him
to
make
stuff
up,
because
he
doesn't
understand
what
really
happened
when
you
took
the
network
out,
it's
an
easy
does
is
an
easy
sell
for
them.
You
know
just
tell
them
look.
This
is
gonna
lessen.
The
pain
is
gonna.
Make
us
more
effective
and
they'll
be
less
anger
over
time.
Right.
That's
not
gonna
happen
overnight,
but
it's
gonna
happen
over
time.
B
Iii
feel
like
you
definitely
need
somebody,
you
know
a
little
bit
forward
thinking
and
that
managerial
role,
because
you
know,
in
my
experience
when
you
know
when
there's
a
when
there's
a
hiccup
when
I
change
those
South.
If
you
haven't
established
automated
practice,
if
you
are
doing
some
testing,
you
know
it's
very
easy.
A
B
A
A
B
B
A
And
it's
you
know
what
I
think
it
actually
bring
I
think
it's
important
I
used
to
say
all
the
time
that
there's
little
engineering
and
big
e
engineering
out
in
the
universe
and
I
think
you
know.
Network
engineering
has
been
in
a
sort
of
little
--mode
for
a
long
time.
We
just
listen
to
whatever
our
vendors
tell
us
and
we
don't
really
engineer
that
much
right
and
I
think
this
practice.
A
You
know
it
because
it's
kind
of
insane
logging
into
a
million
devices
at
the
CLI
making
by
hand
it's
crazy,
so
I
think
it's
an
opportunity
for
us
to
grow.
Our
field,
which
has
been
and
for
I
mean
automation,
is
becoming
more
popular,
but
even
now,
I
go
and
I
talk
to
customers
I'm,
also
on
the
vendor
side,
all
the
time
and
you
still
80%
or
more
organizations
to
do
none.
They
just
have
teams
of
people
hammering
away
at
the
CLI
every
every
week.
B
Yes,
yeah
no
agree,
that's
definitely
definitely
what
I'm
seeing
and
you
know
it's
sometimes
some
days.
It's
really
tough
to
be
optimistic,
but
other
days
you
kind
of
see
what
it
looks
like
when
it's
successful,
and
it's
really
a
breath
of
fresh
air
kind
of
kind
of
gives
me
that
optimism
that
networking
networking
is
evolving
because
there
are
a
number
of
teams,
a
number
of
companies
and
they
are
successful
at
it
and
to
see.
A
Yeah
yeah
agreed
I,
it's
it
isn't.
That's
probably
the
best
part
of
being
an
evangelist
for
automation
is
when
you,
when
you
get
that
you
can
see
the
light
bulb,
go
off
in
their
head
and
and
that's
really
satisfying.
It's
like
hey
it's.
You
know,
I'm,
not
just
yelling,
into
a
void.
You
know
it's
somebody
got
it
and,
and
hopefully
their
life
will
be
better
as
a
result
of
it
yeah
for
sure.
Well,
you
kind
of
all
over
the
place
there
for
topics.
A
So
thank
you
for
coming
on
I'm,
sorry
for
all
that,
I
can't
this
isn't
pretty
incredible
that
we
had
three
things
go
wrong
on
this
dream
in
one:
it
ones
dream.
So
what
I'm
gonna
do
is
I'm
gonna,
edit
it
I'm
gonna,
have
to
edit
this
now
anyways
and
we'll
see
we'll
see
what
happens
there,
but
I
I
just
want
to
apologize
to
everyone
that
it's
pretty.
This
won't
happen
again.
That's
pretty
rare
I'm
gonna
get
a
new
headset.
A
I'm
gonna
get
another,
because
this
is
the
set
this
the
second
time
the
headset
died
on
me
this
time.
If
the
battery
didn't
die,
the
Bluetooth
disconnected-
and
it
makes
a
sound
that
when
it
this
happens,
and
it's
searching
again
for
Bluetooth,
it
just
did
it
for
like
in
the
middle
it
was
connected.
It
was
working
why?
Why
would
it
do
that?
I
just
need
to
get
rid
of
that
headset
and
get
another
one,
and
so
I'll
be
editing
this
together
out
of
we,
we
actually
recorded
it
with
two
different.
A
B
B
A
Okay,
so
you
can
follow
Anthony
at
permit
any
any
on
Twitter
and
I
encourage
you
to
do
so.
I
will
put
a
link
into
his
talk
at
Shina.
It's
on
youtube
about
the
importance
of
testing
and
network
automation
and
I
highly
recommend.
You
watch
it's
not
that
long,
but
it's
it's
very
it's
it's
fantastic
lots
of
great
insight,
a
little
bit
of
humor.
So
please,
please
check
it
out
and
again
I'm
cloud
toad
at
CL
ooh,
you
deed,
Toa
D.
You
can
follow
me
on
Twitter
and
again
the
project
is
entering
labs.