►
From YouTube: CNCF SIG Storage Meeting 2019-12-11
Description
Join us for Kubernetes Forums 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
CNCF SIG Storage Meeting 2019-12-11
A
A
B
B
A
A
C
B
A
A
A
A
A
So
the
first
thing
on
the
agenda
is
which
I'd
like
to
bring
up
is
the
dragon
flying
project,
so
just
as
a
just
as
a
refresher
dragon
fly
presented
to
the
TOC
in
August
and
they
the
TOC
said
that
they
needed
some
input
from
the
from
the
cig
to
help
with
their
with
their
process.
So
so
dragon
flying
is
a
little
unusual.
It's
it's!
It's
a
it's
a
container
container,
registrar
or
registry,
similar
to
to
sort
of
the
Quai
project
and
and
harbor
I
guess
in
some
ways.
A
And
therefore
you
know
there
was
some
confusion
as
to
which
signals
would
actually
fall
into,
but
they,
but
they
passed
the
storage
saying
if
we
could
have
a
look.
The
the
project
presented
to
us
on
the
28th
of
August
this
year
and
in
the
meeting
minutes
you'll
find
a
link
to
the
to
the
recording
of
that
presentation.
A
So
I
was
I
was
hoping
that
anybody
had
some
notes
from
the
presentation
from
back
in
August
and
I
realized
as
being
a
while
now
or
if.
If
you
have
time
to
review
the
recording
and
put
some
notes
down
together,
I
was
just
in
the
process
of
doing
that.
I'm
doing
that
today
and
I'll
show
them
rounds
with
the
with
with
this
team,
but
I
could
really
do
with
one
or
two
other
people
just
just
trying
to
try
to
have
a
look
as
well
and
reminding
themselves.
A
A
I
know
I
at
the
time
I
had
just
done
some
notes,
so
there
are
some
security
concerns
and
some
question
marks
about
consistency
of
files
and
things
like
that
which,
which
I
think
probably
merits,
merits
sort
of
being
brought
up.
You
know,
even
if
the
projecting
sort
of
turns
around
and
says
that
they're
working
on
it-
that's
that's,
probably
a
good
thing,
but
at
the
time
I
recall
that
three
hands
some
some
question
marks
whether
it
should
move
forward
into
incubation
and
I.
D
D
Incubation,
okay,
correct,
but
that's
a
good
point:
by
what
measure
do
we
look
at
that?
Is
it
worth
taking
our
concerns
to
this
security,
sig
I,
don't
know,
I
mean
I,
think
that
process
is
still
being
formulated.
I,
don't
think
it's
a
bad
idea
if
they
have
time
but
I,
don't
know
what
their
backlog
looks
like.
A
Yeah,
so
I
think
you
know
to
the
bar
to
move
to
to
incubating.
Is
you
know?
At
least
we?
We
have
to
feel
that
it's
going
to
adds
value
to
to
the
field
and
I
think
if
there
are
some
some
sort
of
significant
concerns,
either
around
sort
of
things
like
data,
integrity
or
security,
I
think
I
think
it's
works.
A
So
so,
yes,
so
what
I'll
do
is
I'll
put
my
notes
together
in
a
in
a
Google
Doc
and
share
the
link
outs
to
the
to
the
mail
eNOS.
Then,
if
anybody
else
can
then
add
on
to
that
that
doc,
that
would
be
really
appreciated,
and
then
we
can
link
it
back
into
the
into
the
PR
for
the
project
so
that
the
talk
has
the
information.
A
Nope,
okay,
the
the
second
thing
I'd
like
to
I'd,
like
to
just
provide
an
update
to
to
the
rest
of
the
to
the
rest
of
the
team,
is
a
bit
of
an
update
about
some
of
the
process
and
workflow.
That's
that
some
of
the
persons
work
through
discussions
that
we've
been
having
so
and
an
airing.
Please
keep
me
honest
here.
A
A
While
we
were
at
cube
con
and
laser
sports
our
documents
together
to
to
a
minute
some
of
that
meeting
and
following
that
yang
Lee
who's,
the
the
toklas
on
for
the
storage
sake
has
has
said
that
it
would
be
a
good
idea
for
us
to
to
have
regular,
updates
and
regular
discussions
about
some
of
this,
so
that
we
can.
We
can
formalize
some
of
these
processes
and
workflow
for
things
like
doing
project
reviews
and
providing
feedback
into
into
the
TOC
processes.
A
Cool,
oh
hi,
quittin
I,
see
you
just
joined
us
there,
so
Quinton
just
just
for
your
benefit.
We
we
had
a
quick
discussion
about
the
Dragonfly
project
we
were.
We
were.
We've
agreed
that
we're
going
to
collect
some
notes.
I
recall
that
you
had
some
concerns
when
we
had
originally
seeing
the
presentation
of
dragonfly.
So
when
I
circulate
the
notes,
if,
if
you
could
add
to
that
or
annotates,
the
denote
stuff
would
be
useful
sure.
C
I'll
be
glad
to
do
that
very
happy
to
on
that
topic.
As
a
general
topic,
I
was
just
thinking
about
this.
We've
we've
dropped
the
ball
on
a
couple
of
these
things,
and
some
of
them
are
probably
my
fault.
So
this
is
not
not
blame
storming
by
any
means,
but
I
was
wondering
if
we
don't
want
to
elect
a
secretary
as
in
as
in
a
company
secretary
I,
don't
know
what
the
term
in
America
is,
but
I
know
in
Europe.
C
You
know
things
like
filing
regulatory,
stuff
and
financial
statements
and
listen
that
the
next
thing
and
typically
runs
the
board
meetings
and
things
like
that
do
duly
want
to
perhaps
elect
a
secretary
whose
job
it
is
to
make
sure
that
that
we
record
everything
that
we've
promised
to
do
and
that
somebody
you
know,
gets
allocated
to
do
it
and
it
gets
reallocated
if
necessary.
All
that
kind
of
stuff-
oh
I,
mean
if
one
of
the
chairs
wants
to
do
it.
D
I
I,
don't
think
any
of
us
have
time,
but
I
think
it's
needed,
but
I
don't
know
that
it
has
to
necessarily
be
one
person
all
the
time.
I
guess
is
what
I
would
yield
to
I
mean
I
told
Alex
I.
Would
you
know
help
by
being
a
co-lead
rotating
through
creating
the
agenda
and
providing
status
to
the
cig?
Maybe
if
we
could
or
the
tech
leads
could
help
rotating
doing
that?
D
C
And
I
was
testing
a
new
put
an
additional
person,
so
not
not
the
chairs
or
the
cig
leads
I
mean
that
the
problem
is
with
these
rotating
things.
Unless
some
individual
is
kind
of
like
keeping
tab
on
on
who's
actually
doing
these
things,
they
falls
through
the
gaps
which
is
what's
been
happening
and
and
as
I
say,
this
is
not
blaming
anyone
I'm,
probably
as
much
to
blame
as
anyone
else,
but
you
know
everybody's
too
busy
to
do
that.
C
A
I
think
I
think
we
certainly
look
with
dragonfly
with
with
with
well
it
slipped
through
the
cracks
we
moved.
We
moved
on
to
other
tasks
before
finishing
off
that
task.
Basically,
so
I
mean
there
are
two
things:
I
think
we
should
just
become
more
disciplined
eyes
as
a
as
a
as
chairs,
which
is
one
we
should
probably
you
know,
review
the
previous
meetings
minutes
and
make
sure
that
we've
closed
off
things
and
if
we
haven't
keep
them
on
the
current
list.
A
The
the
other
thing
that
I've
done
just
to
keep
things
sort
of
semi
saying
is
at
the
top
of
the
meeting
minutes,
I've
just
added
a
list
of
things
that
are
in
play
and
I
think
we
should
keep
that
list
of
things
which
are
in
play
at
the
top
of
the
meeting
minutes
as
and
just
updated
for
during
four
at
every
meeting,
as
a
simple
way
of
just
not
losing
track
of
things
because
I
think
will
happen
with
dragonfly
is
we
we
just
moved
on
to
other
stuff
forgot
about?
Will
I.
C
Mean
Alex
I
feel
very
guilty
that
you,
you
are
carrying
a
lot
of
the
burden
of
the
cig
in
general.
You're
doing
you
know
some
of
the
work
that
cig
leads,
and
certainly
some
of
the
work
of
the
other
chairs,
including
myself
so
I
mean
conceptually
I'm,
totally
fine,
if
you're,
basically
taking
this
job
as
as
company
secretary,
but
but
I
really
don't
want
to
you
know
over
it.
We
use
your
your
charity
with
regards
all
of
this
work.
A
D
A
D
B
D
A
A
Is
a
bit
slower
than
we
had
hoped,
mostly
because
of
keep
coming
Thanksgiving
and
a
couple
of
other
things
in
the
last
couple
of
weeks,
but
I
just
want
to
do
a
recap
of
the
other
documents
that
we
have
in
progress
right
now.
So
just
going
from
back
to
front
the
think
today
is
dark.
That
sugu
was
updating.
We
had
had
a
fair
amount
for
discussion
and
we've
loaded
up
the
document
with
with
comments
super
I
believe
you
were
on
the
case
to
leave
in
the
comments
into
some
paragraphs
right.
Yes,.
A
All
right,
that's
that's!
That's
really
useful
yeah
I
mean
if
we,
if
we
can,
if
we
can
aim
to
to
so
the
next
sigma
thing,
is
Christmas
Day,
so
which
obviously
means
where
it's
not
happening
so
the
one
after
that
will
be
the
the
one
in
the
first
week
of
January.
So
if
we
could
try
and
target
sort
of
mid-january
for
some
of
these
things,
it
would
be
really
good
similarity
on
the
performance
and
benchmarking
documents.
I've
put
some
work
into
some
of
the
content
around
the
the.
A
A
A
Now
let
me
assist
on
the
call
all
right
so
in
the
use
case,
doc
I
think
we
we
talked
about
last
time.
We
added
in
the
information
to
keep
the
use
cases
somewhat
generic,
but
also
provides
information
for
multiple
projects,
and
we
had
a
bit
of
a
debate
about
how
to
make
sure
that
this
wasn't
seen
as
some
sort
of
CN
CF
endorsement.
A
Those
would
be
the
appropriate
patterns
to
use
to
say
optimize
for
performance
which
woke
the
mice
for
consistency
or
whatever
else,
and
there
may
be
more
than
one
option
and
I
went
through
into
the
realist
into
the
minutes.
Just
to
make
sure
I
was
clear
on
this.
We
agree
that
there
might
be
that
there
that
there
will
be
more
than
one
option,
depending
on
different
optimizations,
that
that
might
apply
for
those
particular
use
cases.
D
Is
Louise
still
leading
that
then
I
mean
whose
responsibility
I'd
like
to
see?
Maybe
the
tech
leads,
and
you
probably
need
to
nominate
another
new
tech
lead
as
well
out
to
help
with
that
I
mean
I
know
what
Lewis
was
starting
with
the
first
one,
but
maybe
we
weary
review
that
and
then
go
on
to
you
know
assigning
out
some
of
the
other
use
cases.
What
do
you
think.
A
Yeah
I
mean
that's,
that's
fine,
I
think
Lewis
was
Lewis
was
exactly
that
that
had
put
a
template
together,
originally
he's
obviously
not
available
today,
but
I.
Don't
think
you
know
I
think
he's
still
he's
still
on
the
case.
We
can
follow
up
with
him
and
on
slack,
but
I.
Think
there
certainly
is.
The
personally
is
the
opportunity,
perhaps
for
another
tech
leads
I'm
saying
if
you
have
time
to
to
put
together
another
example
based
on
the
template,
with
the
document
that
Lewis
has
started,
putting
together
yeah.
C
To
to
answer
your
your
previous
question
about
any
concerns,
or
whatever
I
I
just
had
a
question
which
is,
if
I
understand
that
the
sort
of
plan
correctly
it
is,
you
know,
I
just
had
a
very
brief
look
at
the
minute,
make
sure,
whereas
there
were
such
things
before
the
sifrits
it.
According
to
its
meetings
that
happened,
it
sounds
like
so
it
seems
like
they
were
focused
on
applications,
meaning
you
know,
there's
already
a
decision
being
made
well.
C
C
You
know,
Kafka,
they
usually
have
a
requirement,
for
you
know
a
key
value
store
or
a
document
store
or
an
object,
store
or
something
like
that,
and
then
they
have.
You
know.
The
implementation
of
the
object
store
is
the
one
question
some
of
them
are
distributed.
Some
of
them
are
centralized.
Some
of
them
are,
you
know
higher
durability,
blah
blah
blah
and
then
also
those
things
typically
are
layered
on
top
of
other
pieces
of
storage,
and
so
there
are
further
choices
there.
C
So
I
was
sort
of
imagining
that
that
that
might
be
the
Avenue
that
we
attack
it
from
what
is
your
actual
general
requirement
and
then
what
are
the
various
permutations
of
you
know?
Higher
level
and
lower
level
storage
systems
out
there
approaches
to
storage,
and
then
maybe
some
specific
example
was
that
better
or
worse
for
these
police
was
particularly
space.
What.
A
Make
sense
yeah
so
so
I
think
we
got
ourselves
tied
up
in
knots
last
time
because
we
were
because
you're
right
I
think
there
is.
There
is
a
difference
between
applications
or
use
cases
that
consume
storage
versus
some
of
the
things
that
are
actually
providing
persistence,
also
using
other
types
of
storage
right.
So
on
the
one
hand,
perhaps
you
have
I,
don't
know
something
like
Prometheus
or
something
like
that
CD
or
something
nope.
Let's
see,
these
are
an
example.
A
A
It
could
be
a
key
value
store
whatever
it
is
right
and
then,
on
the
other
hand,
you
have
perhaps
things
like
Mineo,
which
is
an
object
store,
but
there's
probably
consuming
volumes
from
somewhere,
or
perhaps
you
have
a
database
which
is
a
persistent
layer
in
its
own
respect,
but
it's
probably
consuming
storage
again
from
somewhere
and
I.
Think
I
think
we
we
kind
of
got
into
a
really
sort
of
tight
loop
debate
about
this,
because
the
example
that
Louise
put
together
was
about
Mineo
and
therefore
it
was
sort
of
it
looked
like
it.
A
We
were
promoting
Mineo
as
a
storage
system
and
that
that
that
was
sort
of
I
think
the
root
of
someone's
debate.
What
I?
What
I,
really
think
we
we
need
to
do
for
this
to
be
usable,
is
to
take
into
consideration
the
specifics
of
particular
use
cases,
so
so,
for
example,
just
just
to
pick
an
example
if
we
wanted
to
say
Prometheus
say
because
that's
a
CN
CF
project
and
it's
an
it's
a
use
case
that
needs
to
consume
large
amounts
of
storage.
A
So
so
we
should
be
able
to
say
look.
This
requires
a
volume
in
the
fine
system.
You
probably
are
going
to
have
multiple
instances
of
Prometheus.
Therefore
you
need
to
figure
out
how
to
have
volumes
on
different
nodes.
The
the
I/o
profile
of
Prometheus
requires
throughput,
optimization
versus
latency,
optimization
for
an
example,
because
you
have
replicas
and
Prometheus.
You
probably
not
particularly
focused
on
the
strongest
consistency.
A
C
So
that
that's
sort
of
what
I'm
saying
I
guess
the
subtle
difference
between
what
I
was
trying
to
say
and
what
you're
saying
is
you're
focusing
on
prometheus.
What
I'm
saying
is
Prometheus
is
an
example
of
a
metrics
collection
system.
Then
most
of
them
have
similar
requirements.
They
don't
have
strong
consistency,
guarantees
they
do
have
high
throughput
requirements,
etc,
etc.
I
would
have
no
problem
at
all
using
Prometheus
as
a
sort
of
canonical
example
and
saying.
C
Therefore,
here's
an
example
of
a
good
way
to
put
in
a
monitoring
system
for
I
was
Prometheus
on
top
of
XYZ,
but
I.
Don't
think
that
we
should
be
to
get
too
carried
away
with
Prometheus,
it's
sort
of
one.
It
sounds
like
we're:
promoting
Prometheus,
which
is
not
really
our
responsibility
and
two
I
don't
think
we
would
be
communicating
the
generality
of
the
recommendations
that
we're
making,
which
is
that
you
know
these
kind
of
monitoring
systems
typically
need
these
kinds
of
storage
and
I.
C
D
Thought
we
talked
about
this
so
when
we
went
through
Lewis's
thing
that
that
was
the
changes
we
were
gonna
make
is
we
were
gonna,
be
more
generic
and
that
we
would
then
under
the
exhibit.
We
were
gonna,
add
a
section.
That
said
like
examples
of
this
use
case
and
there
we
would
provide
things
like
Prometheus,
so
we
were
gonna
generalize
out
the
use
cases,
because
we
all
agreed
that
it
shouldn't
start
off
with
like
Mineo
on
this
blah
blah
blah
and
then
I
thought
we
already
went
through
this,
so
I
guess
I'm
confused.
C
B
A
Debate
last
time,
I
think
you're
right
we
should.
We
should
start
with
the
generic
and
then
I
think
what
we
agreed
was
that
we
would
do
the
generic,
and
we
would
have
some
examples
of
what
that
generic
use
case
covered
I
would
like
and
I'm
kind
of
open
to
this,
but
I
would
like
to
kind
of
go
one
step
further.
A
So
if,
if
you
imagine
this
to
be
a
directory
structure
in
a
repo
I
would
love
to
have
the
generic,
but
I
will
also
love
that
the
the
actual
project
owners
that
for
that
category
are
able
to
write
a
more
specific
example
that
defines
best
practice.
For
that
specific
example
to
I
would
I
would
love
it
if
it
would
evolve
into
that.
Eventually,
you
know
so,
for
example,
yeah.
So,
for
example,
you
know
if
the
generic
use
case
was
like,
you
know,
load
monitoring
and-
and
these
are
the
things
that
you
would
like
to
do.
A
It
would
also
be
awesome
and
then,
if
we
go
one
step
further
and
say
well
actually,
if
you're
doing
Prometheus,
yes,
a
file
system
optimized
for
troop,
it
is
important,
but
you
should
look
for
a
file
system
that
can
handle
I,
know
these
type
of
block
sizes
or
these
type
of
file
sizes
or
whatever
else.
You
know,
because
they're
they're
optimizing
for
some
specific
things
that
are
specific
to
that
to
that
project
or
they're
not
use
guys
yeah
and.
C
I
can
imagine
you
know
yes
on
a
on
a
5,
node
cluster
might
look
different
and
Prometheus
on
a
$50,000
truster.
Alright,
we
could
have
to
you
know
different
case
studies,
for
those
to
exchange
of
that.
That
sounds
fantastic,
so
so
I
think,
and
you
need
to
do-
is
perhaps,
as
Aaron
pointed
out
earlier,
perhaps
start
sort
of
Charlene
you
step
out
to
the
right
three
and,
and
that
seems
to
imply
you
know,
creating
that
directory
structure
that
you
mentioned.
C
Perhaps
starting
with
you
know
not
the
leaves
of
that
directory
structure,
but
the
the
trunks,
the
the
big,
the
big
use
cases
figure
out
what
there's
our
relational
database
in
that
new
store.
You
know
monitoring
system
document
store
whatever,
whatever
those
things
look
like
at
least
you
know,
dropped
out
that
thing
and
then
start
you
know
electing.
Perhaps
a
tech
lead
for
each
one
or
something
like
that.
So
we've
got
some.
That's.
A
A
What
I'll
do
is
I'll
I'll
share
at
a
a
Google
Doc
after
this
to
the
team,
and
maybe
we
can
start
brainstorming
that
the
things
that
we
that
we
want
for
those
general
use
cases
awesome
sounds
brilliant.
All
right,
that's
fantastic,
so
those
were
the
main
things
we
had
in
the
agenda,
Amy
that
you
have
anything
from
the
CN
CF,
specifically,
if
you're
still
on
you.
E
It's
a
different
mic.
It's
fine
good
morning,
it's
beautiful
yeah,
the
question
that
was
around
Amsterdam
and
if
we
want
to
be
able
to
do
a
storage
day
there.
A
E
D
I
have
a
suggestion
if
I
may
I
found
it
rather
fascinating.
The
Giant
Swing
in
focus
and
like
I,
went
to
see
NS
day
and
it
was
like
very
customer
focus
very
use
case
focus.
You
know
what
people
are
seeing,
what
vendors
are
doing
and
then
I
go
to
the
cube,
storage,
sig
and
and
we're
very
like
high
level,
design
and
I
understand
they
have
different
purposes,
but
I
I
wish.
D
We
could
find
a
way
to
be
able
to
kind
of
meld
those
two
together
where
we
get
to
hear
from
community
and
customers
what's
missing
and
what's
it's
needed
compared
to
what
we
feel
like
we're
driving
in
the
in
the
sig
in
the
community?
So
particularly
just
a
suggestion.
I
don't
know
if
that
can
happen,
but
but.
B
I,
just
I'm
afraid
that
we
won't
have
enough
time
here
for
the
sixth
or
eighth
based
your
face
that
one
day
we
couldn't
really
finish
discussing
all
the
topics
we
want
to
talk
about.
So
if
we
come
behind
them
together,
then
we'll
be
then
having
less
time
time
for
those
type
of
discussions.
So
I
think
what
you
need
to
have
multiple
events,
so
no
more
time,
I
think
so.
A
A
But
so
so
we're
going
to
go
ahead.
You
know
the
sponsors
who
responsible
the
last
three
are
kind
of
getting
together
and
and
we'll
be
getting
going
ahead
and
and
probably
sponsoring
another
Klan
native
storage
day
too,
and
that
focus
is,
is
all
about
sort
of
the
practical
implementations
of
client
native
storage.
So
we
tends
to
get
end-users
together.
Discuss
actual
use
cases
discuss,
take
through
problems,
discuss,
you
know,
various
implementation,
details
and
examples
and
have
demos
and
a
variety
of
other
things.
A
It's
kind
of
a
mix
between
you
know
what
the
sponsor
is
providing,
what
the
what
the
end-users
provides.
So
so
it's
it's.
It's
a
very
pragmatic.
You
know
how
to
type
of
session,
as
Erin
mentioned,
the
cube,
the
cube
sake
face
to
face
meeting
is
obviously
all
about
well
the
technical
design
and
specking
and
scoping
of
features
and
things
like
that
and
is
very,
very
developer,
focused
and
the
question.
That,
therefore,
is
you
know
what
role?
If
any,
do
we
want
the
CNC
f
sig
to
have
in
organizing
something?
A
E
So
I
guess
my
thinking
about
that.
This
is
also
how
much
work
one
of
these
things
is
going
to
be
to
be
able
to
move
it
over
to
the
sig,
whereas
we've
already
got
kind
of
like
the
cloud
native-like
security
day.
Sorry
storage,
they
just
kept
happening
on
its
own
and
figuring
out
ways
to
be
able
to
have
this
group
participate
in
that
without
taking
on
all
the
heavy
lift.
A
A
B
So
Alex
I
noticed
that
this
time
in
the
CI
stay
there's
no
community
session.
Where
previously
we
do
have
like
a
CSI
session
and
I
see,
there
are
discussions
to
rather
have
a
ciencia
for
a
succession
or
a
CSI
session,
but
then
they
got
dropped.
Probably
there
are
just
too
many
contents
already
I
guess
yeah.
A
B
B
B
E
E
A
It
would
be
that
might
be
particularly
useful
with,
with
we've
been
trying
to
get
and
user
feedback
for
for
a
while
with
sort
of
limited
success.
Because
even
though
we've
we
have
put
a
survey
together,
the
end
user
sessions
weren't
particularly
productive.
But
if
we,
if
we
could
get
us
if
the
secret
could
get
a
slot
at
the
end
user
session,
to
talk
about
the
work
that
we've
been
doing
and
to
maybe
gather
any
feedback
or
information
on
what
they
would
like
next.
And
that
would
be
super
useful.
E
F
This
was
like
at
end
of
the
day,
so
it
wasn't
competing
with
people
wanting
to
go
to
talks,
but
that's
another
format
other
than
a
panel
where
you
have
to
find
you
know
unidirectional
speakers
and
I
think
the
interaction
of
a
birds
of
a
feather
where
you
at
this
point
I'd
say
there
ought
to
be
enough.
Kubernetes
users,
some
in
production,
some
thinking
about
getting
there,
that
the
users
of
themselves
could
drive
just
random
conversations
and
have
it
work
out
pretty
well.
F
It's
up
to
us,
I
mean
we'd,
have
to
get
it
on
the
agenda.
I'd
suggest
an
hour
would
be
appropriate.
It
may
be
30
minutes,
not
enough,
really
what
you
do
when
you
open
these
is
you
you
go
around
and
everybody
does
a
one-minute
intro,
and
that
alone
is
gonna,
gobble
up
15
20
minutes,
but
without
some
context
of
who's.
F
You
know
coming
from
where
I
think
it
doesn't
work
as
well,
and
that
introductions
breaks
the
ice
and
gets
people
talking
amongst
each
other,
so
I'm
thinking
an
hour
and
then
I
think
what
you're
likely
to
have
if
it
worked
well,
is
after
your
hours
up,
nothing
keeps
people
from
self-organizing
going
off,
for
you
know
a
meal
or
a
beverage
and
continuing
on
if
they
found
it.
Interesting
I
think
that
that
is.
F
Only
that
here's
another
of
just
brainstorm,
but
we
had
a
few
of
these
go
down
at
cube.
Con
San
Diego
during
lunch
kind
of
unofficial
self-organized,
put
a
couple
signs
out
on
a
couple
tables
in
the
general
lunch
area
and
self
assemble,
and
it's
a
little
less
organized
and
I
I
think
the
only
way
those
actually
were
promoted
were
speakers
announcing
it
during
a
session.
Saying:
hey,
we're
gonna,
have
this
event
during
lunch
and
look
for
the
signs
and
it's
a
little
less
formal,
but
that
might
work
as
well.
C
Yeah
I
attended
a
few
of
those
and
I
found
them
very
useful.
I
mean
quite
often,
people
are
wandering
around
looking
for
someone.
They
recognize
to
talk
to
over
lunch,
and
it's
great
to
just
have
a
big
sign
saying
you
know
if
you're
interested
in
topic,
XYZ
come
and
sit
at
this
table
and
you'll
have
someone
to
talk
to.
F
A
Thought
that
that
sounds
like
a
really
great
all
of
this
sounds
like
a
really
great.
It
is
I
mean
I,
like
the
idea.
First
off
in
the
end
user
session
of
Bureau
of
birds
of
a
feather
session
sounds
brilliant
too,
and
maybe
we
can
get
some
volunteers
from
the
sake
to
to
to
partake
in
some
of
those
lunch
sessions
as
well.
Yeah.
F
I
think,
if
we're
gonna
do
it,
we
should
self-identify
some
people.
We
don't
want
a
hundred
percent
end-users
and
somebody
should
be
a
moderator
yeah.
But
so
maybe
we
nominate
some
people
who
are
gonna,
go
there
to
be
organizers,
but
that
role
is
really
just
to
promote
talks
amongst
the
users.
In
my
opinion,
good.
E
C
C
A
A
A
I,
don't
know
if
that
actually
translates
to
the
CN
CF
helps
with
the
sponsorship
or
anything,
but
because
these
these
things
are
actually
pretty
pricey
to
to
build
it
up,
but
because
the
so
the
my
my
question
is
what
what
do
we
think
the
sake
creates
contributes
to
to
that
storage
day,
because,
actually,
as
it
happens,
a
lot
of
the
sig
members
do
already
participate
in
the
cognitive
storage
thing
in
in
some
form
of
or
in
some
form
or
another.
You.
C
C
C
Yep
and
the
third
one
would
be
to
you
know,
assist
you
know,
so
we
could
people
hear
the
tech
leads
or
whatever
could
chair
panels
will
do
whatever
might
be
useful
and
they
would
actually
be.
You
know
in
their
official
capacity
as
tech
lead
or
chairperson
of
the
CNC,
a
storage
sink.
Maybe
they
do
that
anyway
and
they're
just
maybe
never
had
much
I
don't
know,
and
the
third
one
could
be
organizing.
You
know
helping
with
the
organization,
so
so
choosing
you
know,
maybe
adding
some
value
to
the
content.
A
Right,
no,
that's
all
really
good!
You
know
what
I'll
take
this
to
the
to
the
cannon
for
storage
there,
guys
and.