►
From YouTube: Foundational Infrastructure Working Group , 2023]
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Beyond,
do
you
want
to
do
the
honors?
You
had
a
lot
of
items
or
show
I
share
my
screen.
B
You
can
share,
I
can
start
with
Titans.
Yes,
okay,.
C
B
So
we
should
start
yeah
with
the
RFC
so
about
some
other
people
are
joined
because
of
this.
Yes,
so
I
went
over
the
RFC
and
the
description
looks
good.
So
it's
really
explaining
current
details.
The
two
options
we
have
back
advantages
disadvantages.
B
And
you
know
we
are
getting
a
sleep
yeah
that
it
was
nice
I.
B
B
Yeah
I
think
also
expectations
that
we
take
the
ownership
of
the
contribution
and
it
will
be
easier
for
us
to
maintain
that
one.
D
So
if
the,
if
we
go
the
route
of
the,
what
we
call
pcab
light
integration,
basically
wrapping
TCP
dump,
then
the
expectation
from
us
would
be
that
it's
it
would
be
handed
over
to
Bosch,
because
it's
just
a
very,
very
slim
layer.
Basically,
if
we
had
gone
the
other
way
like
that,
like
pcap
full,
if
you
will
integrate
it,
then
we
would
have
retained
the
the
ownership,
because
it
would
be
basically
just
calling
libraries
that
we
still
own
and
still
deploy.
A
D
We
discussed
it
and
I
think
internally
we
decided
the
the
shared
Parts
would
be
so
small
that
it,
the
overhead
of
having
the
shared
dependency,
wouldn't
be
justified
in
our
opinion,
max
I,
don't
know
if
you
want
to
elaborate
here,
yeah.
E
So
the
the
thing
is
that
the
the
deduplication
and
those
things
are
very
involved
like
they.
E
You
have
a
lot
of
channels
and
things
and
they
tightly
integrate
with
the
grpc
logic
that
we
have
like
reading
from
the
stream
and
so
on,
and
it
would
probably
be
quite
some
effort
to
make
them
shareable
in
that
sense,
so
it
the
sensible
approach
in
my
opinion,
would
be
to
take
over
the
parts
that
are
relevant
and
just
copy
them,
because
we
have
them,
but
we
will
need
to
adjust
them
and
then
we
can
just
maintain
them
separately,
because
the
overhead
I'm
having
a
library
introduce
independencies
could
also
make
it
harder
for
us
to
change
things
internally
in
the
API
which
we
wouldn't
want.
A
Yeah
so
like
the
benefit
of
to
me
of
having
a
shared
library,
is
the
fact
that,
like
the
ownership
part,
is
clear
right
like
that
library
is
maintained
by
the
pcap
team,
because
you
all
are
domain.
Experts
on
that
part
and
the
Porsche
core
team
isn't
per
se
right.
But
they
maintain
all
the
bits
of
like
shipping
or
shipping
the
packets
like
over
SSH,
and
all
that,
like
the
CLI
bits
and
like
just
how
you
are
or.
F
A
You're
gonna
visualize
all
that
the
the
stream
to
data
and
make
sense
of
that
I
think,
like
probably,
there
will
be
no
features
added
from
the
Bosch
core
side,
all
right.
So
if
you're
gonna
evolve,
the
pickup
full
version
further
with
additional
features
and
things
that
might
as
well
I
I,
don't
know
what
the
roadmap
is.
Gonna
look
like
right,
so
I
find
it
difficult
to
figure
that
or.
A
Yeah
like
it
could
be
that
it's
feature
complete
right
and
it's
pretty
simple,
but
it
will
be
Frozen
in
Time
right
like
if
the
pickup
full
thing
develops
further.
Those
changes
could
be
backboarded,
but
could
also
not
happen,
and
you
don't
have
that
if
you
go
to
Shared
Library
approach.
E
Yeah
I
get
your
point.
The
question
that
I
have
in
mind
is:
if
that
specific
piece
that
we
would
share,
will
evolve
at
all.
Currently
I,
don't
really
see
much
room
for
that,
but
then
again
I
don't
know
we
haven't
done
it
yet,
so
we
don't
really
know
but
yeah.
It's
really
just
duplicating
packets,
I
I,
yeah
I
have
a
hard
time
finding
things
that
could
evolve
that
logic
and
would
be.
E
A
E
No,
so
so
the
output
would
just
be
binary,
so
the
the
PCAT
file
format
is
binary,
so
we
would
just
write
that
either
to
stand
it
out
to
a
file
like
like
TCP
time
does,
but
so
tcpnam
also
provides
this
way
of
displaying
packets
printing
them.
That's
something
we
we
don't
do
or
we
don't
have,
so
you
will
always,
but
what
you
could,
for
example,
do
is.
E
F
A
And
that's
also
the
the
interface
with
the
full
pcap
implementation
that
you're
doing
the
same.
E
Yeah,
so
also
for
this
in
the
CI,
we
don't
at
least
I'm
not
aware
currently
have
an
option
to
like
have
the
have
a
similar
print
out
to
to
what
a
TCP
dump
candle
I
think
there's
some
debug
logging
that
just
shows
the
kids
are
riding,
but
it's
not
like
a
valuable
output.
If
you
want
to
look
at
packets.
D
I'd
say
we
we
will
see
how
much
could
be
shared.
What
like,
as
we
go
right
now,
it's
to
us,
at
least
it
looks.
It
looks
like
a
very
tiny
bit.
Yeah
I
mean
if
it
makes
you
comfortable,
and
we
could
think
about
sharing
this.
Just
this
like
the
deduplication
algorithm,
which
we
need
in
in
both
cases,
that's
true,
but
it's
basically
yeah
50
lines
of
code,
or
so
so
it's
very
very
little,
but
that
we
could
think
about
that's
true
yeah.
D
If
everything
else
like
the
like
the
plumbing
of
how
this
would
be
like
put
into
the
boss
CLI
itself,
that's
outside
our
scope.
In
my
opinion,
to
be
honest,
because
that's
just
how
Bosch
works
right,
how
the
Bosch
CLI
works
the
getting
the
context
of
the
SSH
session
and
such
so
that's,
that's,
obviously
not
not
shareable,
because
it's
push
only
yeah.
A
Yeah,
no,
that's
definitely
not
shareable,
and
also
that
shouldn't
change
at
all
right
like
that.
That
should
be
done
once
probably
using
the
example
of
the
the
SSH
streaming
logs
thing
right
that
that
I
Linked
In
the
Beginning.
It
should
be
really
similar
to
that
workflow
and
then
just
calling
out
to
that
potentially
shared
library
or
or
not.
A
Right
like
it's
to
me,
it's
not
a
hard
blocker,
it's
not
a
hard
requirement,
but
it's
a
preference
like
it
feels
like
a
clean
air
contract,
especially
between
the
two
teams
right
like
who
owns
what.
A
It
would
make
it
easier
to
review
as
well
right
because
we
can
see.
Oh,
this
is.
A
F
A
Really
understand
they
need
to
go
and
understand
that's
outside
of
her
domain.
That
makes
it
a
bit
more
involved,
not
undoable.
Yes,
yeah,
nothing
to.
D
B
D
A
Yeah,
that
has
my
preference,
because,
especially
around
like
ownership
right,
we
don't
have
to
have
to
go
to
the
CLI
repo
you
just
can
pull
it
in
and
like
it
doesn't
have
to
be
like
like.
Even
if
there's
gonna
be
like
a
a
breaking
change
to
the
package
like,
we
can
probably
figure
that
one
out
between
the
two
teams
right,
like
don't
think
about
it,
like.
F
A
Can
never
create
a
breaking
change
or
whatever
right
like
we
can
all
be
flexible
about
that,
but
we
will
have
to
work
again
together
again.
At
that
point,.
A
B
Regarding
the
options,
do
some
work
is
additional
comments
or
buzz
it.
We
agree
that
we
go
with
the
right
one.
A
D
D
D
So
so
that
when
you
roll
down
a
little
more,
there's
like
the
two
integrated
pcap
release
with
Bosch
and
add
pickup
light
as
a
thin.
That's
basically
the
two
options.
Yes,
and
when
you
let
they
are,
they
have
links
and
when
you
click
them,
you
will
see
like
the
benefits
and
drawbacks
of.
C
A
B
F
C
Maybe
one
thing
to
highlight,
because
Ruben
you
said
that
you
would
like
to
see
this
using
the
same
mechanism
as
the
log
retriever
thingy.
This
might
be
slightly
different,
because
our
idea
was
then
to
use
an
SSH,
Master
mode
control
connection
that
can
use
multiple
channels,
so
one
would
be
using.
This
will
be
called
actually
calling
TCP
dump
and
would
retrieve
the
data
that
we
get
and
the
other
channel
could
be
used
to
send
signals
to
that
process.
C
Exactly
I
mean
conception
is
still
the
same,
and
this
is
also
something
that
can
be
done
using
the
go
SSH
the
standard,
Library
code,
SSH
thing.
B
B
C
So
the
the
main
idea
for
why
we
wanted
to
just
integrate
pickup
release,
as
is,
was
to
give
it
more
surface
area
and
get
more
feedback
and
improve
on
the
functionality
on
the
robustness
of
it
with
a
wider
audience.
Basically-
and
it
gives
you
a
bit
more
because
it
has
a
central
component
that
is
overall
slightly
more
complex,
it
gives
you
a
bit
more
flexibility
for
the
future.
So,
for
example,
the
things
that
Dominic
mentioned
in
passing,
with
maybe
dumping
data
to
S3
and
so
on.
C
So
all
of
these
kind
of
asynchronous
things
where
you
don't
necessarily
stream
to
your
client
could
be
done
later,
with
the
pickup
release,
the
full
version,
so
to
say
they
could
not
be
done
easily
with
the
SSH
version,
because
that's
it
is
a
solution
for
this
exact
use
case,
but
for
nothing
else,
pretty
much
yeah.
That's
really
the
kind
of
trade-off
that
we
have
now
to.
F
C
D
Yeah,
we
don't
have
two
strong
feelings
because
you
don't
have
to
like
you
can
have
both.
You
can
still
do
like
the
light
version
and
still
also
deploy
the
full
box
release
on
top,
and
you
don't
have
any
problems,
because
the
light
version
Don't
doesn't
deploy
any
agents,
it's
just
like
doing,
SSH
and
calling
CSP
dump
and
the
full
version
doesn't
even
call
a
TCP
dump.
So
it's
totally
a
totally
separate
worlds.
D
So
that's
why
we
don't
really
mind
dropping
the
full
version
as
integrated
with
part
itself,
but
we
will
probably
still
deploy
it
in
our
production
because
we
just
might
need
more
features,
but
for
the
community
I
think
it's
perfectly
fine
to
have
the
light
version
like
as
a
default
thing.
You
could
still
then
opt
to
add
the
full
version
on
top.
If
you
wanted
I.
A
F
C
Maybe
adding
to
what
Dominic
said
that
is
basically
option:
zero,
not
doing
anything
in
this
RFC,
not
integrating
with
Bosch
just
using
pickup
release,
as
is
because
pickup
release
itself
has
a
Bosch
capturing
mode
that
Still
Remains
as
it
remains
untouched.
For
now,
it
was
just
the
idea
that
maybe
we
could
make
it
more
streamlined
overall,
if
we
merge
that
with
Bosch.
This
is
this
option,
one
thing,
but
it
seems
maybe
more
complex
than
it
needs
to
be
for
this
first
step.
B
And
one
everyone
was
showing
the
drawbacks
of
the
White
Version
yeah.
The
discussions.
I
understood
that
also
performance
is
a
I
think.
So
we
have
a
small
performance
or
throughput
with
the
SEC
wide
version.
Is
this.
E
Well,
to
be
honest,
we
don't
really
know
okay,
so
we
know
the
overhead
of
the
Geocache
solution,
but
we
have
a
tested,
the
SSH
solution
for
obvious
reasons,
so
that
would
be
something
that
we
could
also
look
into
set,
even
if
this
turns
out
to
be
a
low
performance
version
that
still
can
serve
as
a
like
Bare,
Bones,
debugging
tool.
You
could
have
the
full
solution
when
you
need
more
performance
or
whatever,
but
that's
something
that
we
have
to
explore
once
this
is
actually
built.
D
Yeah
there's
some
minor
thing:
that's
technically,
we
are
depending
on
an
external
binary,
TCP
dump
that
has
to
exist
on
the
stem
cell,
I
mean
obviously
it
does,
but
like
it
could
be
the
version,
changes
and
there's
different
flags
or
whatever
that
we
don't
know.
So
it's
not
a
batteries
included
approach
anymore
and
we
bring
all
the
binaries
but
I
think
it's
a
very,
very
minor
drawback,
because
TCB
dam
is
very
stable,
I'm,
not
sure,
doesn't
change,
usually.
D
It
could
be
that
it's
like
no
longer
shipped
by
default
and
it's
just
missing
like
Ubuntu.
They
come
up
with
some
new
ideas
every
week
now
and
that
could
break
basically
the
feature.
Of
course
we
would
have
tests
for
it,
but
yeah.
A
A
F
D
Oh
yeah,
it's
the
wrong
version.
It
doesn't
yeah
I,
don't
think
we
have
changes
in
Bionic,
I'm.
B
A
A
Really,
really
old
environments
by
just
updating
your
your
boss,
CLI,
you
will
be
able
to
use
this,
which
I.
A
Will
review
it
as
well,
but
like
let's,
let's
try
to
aim
for
like
next
Toc
meeting.
We
try
to
start
the
final
comment
period.
B
B
B
And
afterwards,
no
big
changes
can
happen.
Oh
changes
which
impact
the
decisions,
but
we
have
an
agreement.
It's
I,
oh
checked
it
and
I.
It's
I,
don't
have
anything
to
want
to
come.
B
B
We
don't
have
any
other
symptom
with
multiple
options.
Maybe
we
can
put
it
the
beginning
problem
summary
yeah.
A
C
F
A
C
And
also
these
pros
and
cons
at
the
end,
this
comparison
also.
C
F
D
A
Would
be
added
before
the
before
next
week
before
we
start
the
final
comment
period.
C
Okay,
so
we'll
wait
for
your
review
and
your
thumbs
up
we're
down
and
if
there's
anything
that
you'd
like
to
still
see,
we
can
add
that,
and
otherwise,
we'll
add
this
note
that
for
the
time.
A
B
B
And
the
branch
protection,
so
we
had
some
additional
information
here,
but
regarding
the
branch
protection.
A
I
think
we'd
like
need
to
figure
out
I
I,
agree
with
Joseph's
point
a
bit
but
I
believe
there's
also
these
new
keys
because,
like
the
the
normal,
what
was
it?
You
have
those
fine-grained
access,
Keys
now
right
right,
where.
C
B
A
B
B
And
most
probably
not
of
those
in
our
Superior
Bosch,
so
we
have
to
talk
to
that
right.
Yes,.
A
B
B
Well,
yeah
240
I
talk
also
to
the
performance
of
the
time
to
assume
is.
A
B
Yeah,
okay:
we
can
offer
to
sync
with
the
software
yeah.
A
Yeah,
yes,
this
is
something
I
wanted
to
discuss
and
forgot.
Last
week
and
Roman
pointed
out,
Matthias
has
been
doing
a
lot
of
reviews.
I
was
wondering
if
he
has
already
I,
don't
know,
get
out
the
required
contributions
to
actually
be
eligible
for
approver
status.
A
I,
don't
have
an
answer.
I
haven't
looked
that's
more
of
a
maybe
it
makes
sense
to
check,
but
there's
no
official
number
right.
That.
A
E
A
I
think
something
somewhere:
okay,
yeah
here
approval
requirements,
there's
an
RFC
problem
proposal,
blah
blah
blah
blah
blah
this
yeah,
okay,
very
good
reviewer
for
at
least
three
months.
Oh
20
yeah
should
be
easy,
I,
don't
know,
I,
don't
know.
If
this,
but
I
mean
like
you
could
look
at
at
least
look
into
how
far
you
are
well
and
if
you
haven't
met
this,
you
can
create
a
PR
to
promote
yourself
and
just
submit
the
proof,
or
let
me
continue
that.
A
B
About
toncho,
okay,.
A
B
And
I
don't
yeah
I
can
check
this
in
just
the
boss.
A
Almost
done
people
seem
to
be
on
professional
about
looking
at
their
trying
to
PR
reviews.
Dude
I
just
give
my
okay.
B
A
D
A
A
F
E
A
A
Want
me
to
approve
it
as
well:
no
someone
else
we
look
at
it.
I
think
that's
better
for
this
one,
the
director
well
so.
A
See
I
think
Joseph
is
working.
Please.
A
A
C
A
A
It's
really
difficult
to
get
a
GPU
GPU
support
in
one
of
the
best
clusters
that
we
have
available
to
us,
and
the
next
thing
we're
going
to
do
is
Kuda.
F
A
This
is
about
I,
think
long
had
an
end-to-end
thing
where
he
was
able
to
push
an
application
to
CF
that
did
an
AR
I,
don't
know,
had
a
AI
model
to
generate
some
images
or
something
he
was
able
to
run
that
on
CF.
He
had
to
do
like
lots
of
patches
in
in
different
places
to
to
have
this
show
up
at
the
container
level
and
I
think
this
is
an
effort
of
trying
to
get
these
things.
Upstream.
A
Yeah,
so
you
need
drivers
in
the
stem
cell
as
well.
So
this
is
just
about
like
passing
through
the
options
to
the
VM
creation.
Api
call
to
do
the
make
the
hardware
or
expose
the
underlying
Hardware
in
the
VM
yep.
A
So
this
is
I,
think
Connie
wants
to
re-review
and
then
we
want
this
to.
A
A
A
Wait:
we
just
looked
at
this
I'm
going
in
circles,
so
this
one
we
looked
at
that
one.
This
we
skipped
I
believe
it's!
No.
We
skipped
it
because
the
other
thing
needs
to
be
in
reality.
E
F
E
B
A
A
Yeah,
but
it's
in
there
in
the
documentation
right,
you
do
go,
go
Arc
and
Goose
go
OS
Linux!
You
just
put
those
thing
in
right.
A
A
B
A
lot
of
people
with
opinion
in
the
meeting
thanks,
but
maybe
someone
I
my
answer
is-
can
just
try
it
so
another
I
think
materials
in
the
posting.
B
Yeah
so
that
we
have
two
people,
yeah
yeah,
trying
this
and
then
we
can
decide
yeah.
So.
G
Last
time
I
remember,
I,
tried
base,
and
so
inside
the
next
stem
cell
Builder
report,
and
there
is
one
user
read
B
file.
There
is
one
way
to
build
a
md64
architecture
agent
via
so
M1
laptop
via
the
docker
image.
I
tried
it
so
it
for
the
agent
Bosch
agent
release.
It
works
fine
for
the
binary,
but
for
the
Linux
and
it
doesn't
work.
Well,
then
I
switched
to
the
to
switch
to
the
internship
laptop.
A
Yeah
I'm
still
not
switched
over
and
I'm,
also
not
in.
F
A
Hurry
still
on
the
Intel
yeah,
so
if
someone
else
could
confirm
then
like
maybe
leave
a
note
about
that,
it
has
been
tested
on
an
M1
Mac
and
what
you
need
to
do
to
make
that
work.
I
think
then
this
could
be
merged.
F
F
C
A
A
A
And
they
put
out
I
believe
they
just
put
that
in
the
Reds
file
and
Bosch
will
just
use
it.
No
then
any
well.
They
need
to
change
everything
and
I
I
thought.
First,
when
I
understood.
A
He
wants
to
have
just
that
extra
their
company
certificate
in
there
and
then,
if
you
use
like,
if
you
change
like
the
default
CA,
you
provide
that
and
not
the
rest
right.
Then
it
will
generate
the
leaves
using
the
provided.
One
yeah
true.
B
A
Provide
default
CA,
so
they
provide.
You
know:
companies,
CA
certificate
and
private
key
as
default
CA,
and
then
I
think
it
will
be
used
to
generate
the
risk.
E
B
C
A
A
But
it's
is
there
anything
else
people
want
to
discuss
because
we
are
out
of
time
as
well.
F
A
Yeah,
my
when
are
we
going
to
move
some
certain
things
to
the
open
source
pipeline?
Oh
yeah,
we
discussed
that
and
I
think
you
also
have
created
some
stories
in
the
direct
in
the
Bosch,
the
VMware
Bosch
backlog,
so
I
don't
know
where
they
are
yeah.
A
Idea
is
to
at
least
move
the
director
over
yeah
and
the
web.
Io
is
also
like
it's
also
like
a
good
qualifier,
because
it
doesn't
which
one
the
the
bush
bush
IO
Bush
IO
yeah
yeah.
So
all
the
brochures
things
that
whole
pipeline
can
just
I
think
it
can
be
just
one-on-one
because
it
only
needs
like
two
three
credentials.
So.
A
I
think
the
issue
there
is
that
the
it's
hosted
on
a
private
CF
insta
or
a
test
Cloud
Foundry
instance:
that's
hosted
internally
by
vmwareit
the
firewall
and
then
there's
like
a
hole
punched
in
the
firewall
to
expose
it
to
the
outside
group,
but
I
believe
that
the
CF
API
endpoint
is
only
accessible
from
in
inside
VMware,
so
that
one
would
be
a
bit
more
problematic,
but
at
least
the
pipelines
for
all
the
Bosch
release
stuff
right,
like
creating
other
Bosch
releases
or
adding
those
those
could
be
moved
over,
because
those
are
just
updating
repositories.
A
B
F
A
Yeah,
no,
we.
B
B
Yes
and
a
new
one
about
recreating
credentials
rotating
credentials,
yeah,
yes,.
A
That
one
I
the
update
strategy
for
variables
that
that's,
if
you
want
to
go
out
okay,
then
I
will
share
my
screen
for
a
bit
to
actually
have
that
as
well.
A
But
yeah,
if
this
turns
out
to
be
a
really
useful
thing,
we
could
go
with
more,
like
even
smarter
strategies,
basically.
E
My
point
was
only
that
we
do
not
that
that
we
just
used
the
right
data
type
for
the
parameter,
so
that
we
can
accept
it
in.
A
A
The
good
point:
let's
wait
for
Brian.