►
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
Hello
humans
welcome
to
another
edition
of
this
week
in
Cassandra,
and
so
we
are
excited
we
we
actually
three
of
us
on
this
call.
Just
got
back
from
NGC,
see
the
next
generation
Cassandra
conference
in
austin,
so
we've
got
some
cool
stuff
to
talk
about
around
n
GCC,
and
so
this
week
we
have
Patrick
McFadden,
who
I'm
sure
everyone
knows,
chief
evangelist
works
for
datastax
and
we
also
have
eric
stevens
from
protect
wise
joining
us,
and
it's
kind
of
funny.
Eric
and
I
are
actually
both
in
Denver.
A
I
could
have
probably
walked
down
to
his
office,
and
we
could
have
done
this
this
together
from
his
office,
but
I
didn't
think
about
it.
So
here
we
are
in
separate
rooms,
the
wonderful
wonderful
wonders
of
google
hangouts
so
like
we
always
do
if
you've
tuned
in
before
we're
going
to
kind
of
walk
through
a
couple
of
the
blog
posts
that
are
that
are
on
planet
Cassandra.
A
So
if
you're,
following
along
and
the
blog
posts,
what
kind
of
talk
about
a
couple
of
those
and
then
we'll
talk
a
little
bit
about
a
little
bit
about
NGC,
see
and
talk
to
Eric
kind
of
get
his
impressions
that
kind
of
thing
of
what's
going
on
there.
So
if
you're,
following
along
on
planet
Cassandra,
there
were
actually
two
two
links
this
week
that
were
videos
as
opposed
to
sort
of
straight
up
blog
posts.
A
One
is
a
video
of
one
of
the
engineers
at
uber
and
who
also
happens
to
be
like
an
author
of
the
Google
Borg
paper,
which
is
pretty
cool
talking
about
using
apache
cassandra
and
nae
sews.
So
if
you
are
thinking
about
running
Cassandra,
I'm
asos
definitely
check
that
out.
We
kind
of
wanted
to
give
a
quick
shout
out
to
that
just
kind
of
call
that
out
to
your
attention,
yeah
definitely
kind
of
cool.
The
other
video,
though,
that
we
had
on
there
was
Patrick.
You
were
on
floss
weekly,
which
is
a
pretty
big
deal.
B
Know
that
was
really
cool
cast
yeah
podcast
in
a
video
and
it's
in
the
twit
studios
in
downtown
beautiful
Petaluma
California,
the
Liao's
one
last
people
there
to
be
interviewed.
They
were
the
movers
were
moving
things
out
as
I'm
speaking
developed
care,
not
bad.
I
mean
they've
been
there
for
11
years
and
I
was
the
last
guy.
So
I
guess
they
went
out
with
a
bang,
but
it
was
really
good
real
good
talk.
We
we
had
some
confusion
at
the
the
beginning
about
whether
or
not
will
they
get.
B
A
B
It
was,
it
was
I
actually
thanked
him.
I
said
you
know.
This
is
great,
because
this
is
exactly
why
I'm
gainfully
employed,
because
this
is
still
a
problem.
I
got
to
talk
about
some
background
errata.
I
did
call
accumulate
dynamo
based
DB.
It
is
not.
It
is
a
fully
consistent,
each
base
version
database,
and
so
that's
I
did
same
thing
so
Mia
culpa.
B
B
C
B
A
B
Figure
flaws
weekly,
you
get
it,
they
get
a
different
level,
but
you're
right,
it's
eric
nailed,
it
defies
description
and
but
then
again
I'm
sure
will
be
when
relational
databases
came
out.
Trying
to
I
remember
trying
to
explain
that
to
people
like
well,
there's
tables
and
you
connect
them,
and
how
do
you
connect
them
with
wire?
No,
no!
No.
A
Cool
cool
all
right,
so
if
you
are
interested,
definitely
go
check
out
it's
about
an
hour
long
about
an
hour-long
podcast
/
video
that
Patrick
was
on
floss
weekly,
so
find
a
link
on
Planet
Cassandra,
so
switching
gears
a
little
bit
the
other
blog
post
that
we
kind
of
wanted
to
call
out
a
little
bit.
Sebastian
Estevez
who
works
for
Davis
tax.
He
wrote
a
blog
post
about
Cassandra
collections
and
tombstones,
so
another
blog
post
about
tombstones,
which
is
actually
kind
of
a
nice
segue
to
Eric
equipments
were
done
talking
about
in
stones.
It's
not.
A
Love
those
tubes,
so
I
mean
my
takeaway
from
this
was
was
kind
of
that,
like
you
know,
some
two
stones
get
generated
sometimes
in
ways
that
might
not
be
intuitive
to
like
the
average
user
kind
of
thing
is
that,
like,
is
that
an
accurate
description
of
kind
of
what
he's?
What
he's
talking
about
in
this
post?
You
would
have
comments
I'll,.
C
Yeah
tombstones
streams
owners
are
have
many
surprising
characteristics
to
users
they.
They
seem
like
such
a
easy
to
reason
about
thing
on
the
surface,
but
there's
actually
a
lot
of
complexity
in
the
implications
of
different
aspects
of
tombstones,
and
it's
something
that
if
you
really
want
to
get
into
the
details
of
it,
you
don't
enter
into
lightly
or
it's
not
for
the
faint
of
heart.
You
know,
but
there's
there's
news
cases
surprisingly
order
tombstones
get
created
and
people
well,
you
know
they
look
at
their
CF
stats
mayor.
They
can't
believe
like
how
are
their
tombstones.
C
B
It's
really
the
thing
that
the
takeaway
from
this
article
is
working
with
collections,
of
course,
but
it's
just
a
good
lesson
to
understand
the
ramifications.
You
know
every
feature
has
a
upside
downside
and
knowing
what
those
are
always
important.
Things
like
lightweight
transactions,
sound
great,
but
they
have
a
cost
and
updating
a
collection
or
inserting
into
a
collection.
It's
not
that
onerous.
It's
just
you
got
to
know
some
things.
It's
a
part
of
your
data
model.
B
A
C
The
classic
anti-pattern,
and
even
even
in
really
experienced
guys
still
I
still
think
is
this.
Is
that
that
magic
thing
that
really
really
experienced
guys
still
think
if
I
only
just
read
about
this
correctly,
I
can
I'm
sure
I
can
solve
it
this
time
and
and
they're
just
always,
there's
there's
always
some
problem
with
it.
That
is
makes
it.
That's
not
obvious
at
first
yeah.
B
C
B
C
B
A
So
I
mentioned
I
mentioned
at
the
top
Eric.
You
were
based
here
in
Denver
same
as
soon
as
I.
Am
why
you
kind
of
switch
gears
here
and
we'll
talk
about
in
GCC
and
kind
of
talk
about
what
you've
been
working
on
just
start
by
telling
us
like
a
little
bit
about
you
and
what
you
do
and
what
what
protect
wise
does
sure.
C
So
to
me
personally,
I'm
the
principal
architect
here
at
protect
wise
I'm
day,
one
employee,
another
guy
and
I
fight
for
who
has
lower
employee
ID
sins
people
that
started
in
the
same
day.
You
know
the
protect
wise
is
about
a
three
year
old
startup
we've
been
operating
estelle
for
the
last
two
years,
we're
in
network
security
space,
and
we
are
big
data
with
capital
B.
We
do
network
full
packet
capture
and
we
ship
it
to
the
cloud
and
we
store
it
indefinitely.
C
We
store
it
for
as
long
as
you
pay
us
a
story,
and
so
we
were
talking
about.
You
know
petabytes
of
data,
that
we
transfer
up
to
the
cloud
and
store
and
analyze
and
cassandra
is
a
big
big
part
of
that
cassandra.
Is
our
principal
data
store
for
metadata
about
the
packet
data
and
we
stick
a
you
know
well,
north
of
a
billion
records
of
the
Incas
on
here
today.
So
well,.
A
Are
you
guys,
so
are
you
guys
doing
like
as
far
as
things
things
that
you
might
use
to
analyze,
that
data
like
what
do
you?
What,
if
what
kind
of
tools
are
you
using
on
top
of
maybe
your
Cassandra
data
or
even
I,
don't
know
what
you
do
with
the
cold
data
I'm
guessing
you
don't
keep
it
in
Cassandra
forever,
yeah.
C
That's
true:
we
keep
it
on
Cassandra
for
a
while
some
day
lives
in
Cassandra,
essentially
forever.
Today
we
do
for
for
cost
efficiency
reason.
It's
always
cheaper
to
store
data
in
sp.
If
you
can
get
away
with
it,
there's
no
database
in
the
world
that's
cheaper
than
s3.
If
you
somehow
meant
treat
s3
its
database,
we
we
figured
some
things
out
in
that
space
and
it's
not
just
Hadoop
on
s3.
So
here's
a
plug
for
the
summit
come
to
the
summit
this
year.
We're
going
to
talk
about
it.
It's
really
cool,
it
is
really
impressive.
C
I
didn't
have
a
hand
in
that.
So
it's
not
me
toot,
my
own
horn
there,
but
some
of
my
esteem
co-workers
have
done
a
fantastic
job
in
that
space,
but
you
know
the
nice
thing
about
a
sweetie.
Of
course,
if
you
don't
pay
for
cpu
and
process
and
memory
time
for
the
data
at
rest,
but
we
again
it's
always
a
performance
trade-off,
of
course.
So
what
we
do?
We
do
keep
data
around
in
Cassandra
for
a
while.
C
We
we
believe
that
that
we
can
keep
a
really
massive
amount
of
data
and
we
can
store
and
process
and
analyze
its
really
massive
amount
of
data.
That's
sort
of
what
we're
founded
on
and
could
you
know
like
I,
said
cuz
I'm,
just
a
big
part
of
that
spark
is
a
big
part
of
that
DSE
Solar's,
a
big
part
of
that
as
well
yep.
C
A
B
B
I
should
because
I
get
a
lot
of
like
well
there's
another
Cassandra
conference.
How
come
I
didn't
get
out
it?
It's
not
really
a
cassander
conference
it
if
you
hang
out
on
the
dev
list,
which
is
people
who
care
about,
building
and
running
and
not
running,
building
and
in
creating
things
for
Cassandra
itself
like
this
is
all
inside
of
Cassandra,
not
the
user
base,
where
you
know:
hey
user
land.
Let's
talk
about
how
you
use
Cassandra.
B
So
it
is
it's
a
it's
a
conference
where
we
we
dated
stacks
are
sponsoring
it,
but
what
we
want
is
as
a
community
to
get
together
of
those
people
who
care
about
building
Cassandra
and
it's
a
very
small
thing,
small
affair
40
to
50
people.
It
covers
a
large
cross-section
of
our
developer
base.
So,
for
instance,
Eric
was
there
because
he's
doing
some
very
interesting
work
on
compaction
and
there
are
some
people
from
Instagram
that
are
doing
some
very
interesting
work
on
proxies
and
garbage
collection.
B
There
was
some
group:
there
are
some
people
who
Apple
them
there.
They
could
tell
you
I,
can't
tell
you
what
they
talked
about,
because
it's
something
I
can't
talk
about,
but
it
just
lots
of
different
groups
and
it's
not
just
data
sex
people.
It's
it's
just
people
from
the
community.
We
had
folks
from
strada
Oh,
which
was
great,
which
they
do
the
leucine
leucine
index
for
instead
of
secondary
indexes.
B
A
It
was
pretty
cool
this
year,
I
mean
we
had
Microsoft,
some
guys
from
Microsoft
very
odd
yahoo,
japan.
We
actually
had
come
over
from
from
tokyo,
like
that.
I
had
met
on
my
trip
to
tokyo
back
in
march,
which
was
pretty
cool
to
see
them
here
in
the
US
and
even
gave
a
gave,
a
talk
which
was
which
was
pretty
awesome.
So
this
is
basically
a
GCC
is
like
a
get-together
of
committers
and
the
developers
that
actually
develop
Cassandra
features.
A
You
know
the
stuff
that
comes
out
in
santo
releases
that
the
user
base
then
goes
on
to
use
and
so
Eric
you
gave
a
talk
at
n,
GCC
and
I
think
it
was
recorded
I'm,
not
a
hundred
percent
sure.
If
anybody
is
in
listening
and
interested,
we
will
have
a
link
on
the
planet.
Cassandra
blog
post
there
that
has
a
playlist
of
all
the
videos
and
forgive
the
quality.
We
had
a
webcam
setup
kind
of
at
the
last
minute
to
import
some
of
the
talks
that
happened
at
the
a-10
GCC.
A
So
if
you
kinda
are
interested
in
some
of
the
talks
that
happen
there,
you
can
definitely
go
check.
Those
videos
out,
but
so
I,
don't
want
you
to
like.
We
don't
need
to
like
give
away.
You
know
everything
that
you're,
because
I
know
you're
gonna,
maybe
speak
at
summit
about
some
of
this
work.
You've
been
doing,
but
maybe
give
us
like
a
high
level
or
like
a
tease
of
kind
of
like
what
you've
been
working
on
around
compaction
and
and
what
you
talked
about
it
in
GCC,
yeah.
C
If
you
are
interested
in
Cassandra,
the
summit
is
the
most
valuable
thing
you
can
do
with
your
time
in
the
entire
year
and
then
now
we
begin
the
talk
there.
I
would
love
to
see
you
guys
there
and
we
thought
about
the
details
of
that
and
we'll
be
releasing
the
source
code.
You
know
later
this
year,
early
preview
for
people
who
are
interested,
but
we
don't
consider
a
final
yet.
But
it's
but
it's
past
all
of
our
rudimentary
testing
and
we're
really
really
happy
with
the
progress
that
we're
making
with
its
home
very.
A
B
C
There
there
were
a
lot
of
really
good
talks
honestly
by
the
end
of
the
day.
You
know
the
first
day
is
all
talks.
The
second
day
is
a
young
conference
by
the
end
of
the
day,
there's
so
many
incredibly
sophisticated
incredibly
to
elgin
people
that
are
talking
that
you
sort
of
get
exhausted
well
well
before
the
day
is
over,
but
you
know
some
of
the
stuff
that
really
caught
my
attention
was
the
time
series
data
support.
That's
coming
that
is
going
to
be
a
life
changer
for
for
a
lot
of
people.
C
I
think
we
do
a
lot
of
that
same
stuff
in
application
space
and
some
of
those
computations
are
really
difficult
to
do.
When
you
win,
you
can
only
keep
a
limited
amount
of
data
in
memory
because
because
of
your
data
volume
and
so
forth,
and
it's
up
that
it's
much
better
supported
in
a
database
base
than
it
is
in
the
real-time
streaming
space,
so
very,
very
excited
about
that
I.
Think
that's!
A
community
effort
is
that
right,
Pat
Patrick.
B
It
is
there's
well
time.
Series
on
cassandra
has
been
a
thing
since
I
can
remember
because
well,
of
course,
I
do
a
lot
of
talks
on
doing
time.
Series
data
on
Cassandra,
but
it
does
take
a
bit
of
work
on
the
data
model,
definitely
takes
a
little
bit
of
work
on
the
application
side.
There's
no
shortcut
and
like
you
mentioned
this,
is
exciting
because
it
does
shortcut
a
lot
of
the
things
that
need
to
happen.
B
I've
been
involved
in
some
of
these
JIRA's
I
know
that
there
are
some
you'll
see
some
some
names
popping
up
there
somewhat
familiar
like
Jeff
jirsa,
there's
a
lot
of
conversations,
and
probably
more
importantly,
this
is
a
summation
of
many
conversations
that
have
happened
over
the
years.
Things
like
duration
units,
like
I,
would
like
to
look
at
all
my
data
from
now.
B
Until
yesterday
that's
a
query,
but
how
would
you
do
that
now,
given
cql
and
how
it
does
queries
you
couldn't
so,
unless
you're
very
specific
about
your
date,
you'd
have
to
throw
that
logic
and
application.
So
I
know
it
sounds
somewhat
obvious
if
you
used
any
kind
of
time
series
type
database
or
time
series
type
operations
than
a
database,
but
bringing
into
cassandra
is
oh
man,
it's
gonna,
be
so.
C
Sorry,
just
to
expand
I
feature
a
little
bit
will
essentially,
what
it
is
is
the
ability
to
do
pre,
aggregated
data
counts
and
sums,
and
averages
and
divide
two
fields
and
run
functions
on
things
and
floor
timestamp
and
other
things
like
that
and
maintain
running
aggregate
view
of
data
that
your
application
doesn't
matter
maintain
and
that
you
can
select
back
out
as
if
it
was
just
data
that
you
had
written
in
the
first
place.
Yep.
A
C
A
A
B
Yeah
this
DMX,
esta
Diana
fire
from
my
anti-gm
XO
means
love
you
that
all
right,
so
Eric
Eric
stock,
definitely
is
on
my
top
but
and
I'm,
not
just
because
you're
on,
because
I
think
that's
that's
one
of
those
compaction
strategy,
type
things
it
just
makes
sense,
but
the
one
I
really
have
to
say
I
loved
the
most
was
make
the
calls
talk
on
just
this
call
to
action
around
working
with
Cassandra
and
operators
and
I
do
believe.
That
is
a
missing
piece
of
the
puzzle
that
we're
really
as
a
community.
B
We've
avoided,
maybe
not
community
but
has
been
avoided.
There's
some
really
bad
things
in
there
for
operators,
and
it
makes
her
blood
is
miserable
and
his
talk.
If
you
get
a
chance
to
watch
it,
you'll,
probably
laugh
and
cry
same
time,
because
there's
just
pretty
obvious
stuff
in
there,
so
I
was
excited
without
that,
because
that
means
that
we
may
have
some
really
good
positive
action
out
of
it.
A
So
I
have
to
say
just
answering
my
own
question
like
I,
want
to
give
a
shout
out
two
option.
His
talk
on
the
bite
ordered
partitioner
like
totally
like,
resurrecting
it
like
totally
like
blew
my
mind
like
I
was
like,
like
you
know
after,
and
then
they
followed
up
the
next
day
on
the
unconference
day,
where
they
were
trying
to
hammer
out
the
details
like
never
really
thought
of
it.
That
way,
and
sometimes
top
shin
just
comes
up
with
these
things.
A
I
don't
know
if
I
don't
know
if
it's
actually
going
to
be
a
legit
thing
like
a
legit
option
to
bring
back
the
bite,
ordered
partitioner
in
some
sort
of
smarter
form
that
uses
sort
of
heuristics
to
kind
of
kind
of
a
maximized
locality.
But
it's
an
interesting
problem
to
solve
and
and
yeah
I
just
is.
It
was
only
like
a
it
was
supposed
to
be
a
five-minute
talk
and
I
do
have
to
say
a
ten
GCC
like
all
the
Lightning
talks,
maybe
Eric's,
excluded
I,
think
you
probably
and
Patrick
you
did
too.
B
A
B
Holly,
it's
it's
a
pretty
interesting
concept.
That
was
another
thing
that
was
discussed
amongst
the
they're,
the
ranks
there
but
is
composable.
You
know
we
have
baked
consistency,
levels,
one
I'm,
local
Gorham
whatever
and
his
his
idea
was
making
it
so
that,
given
a
topology
and
the
snitch
involved,
you
should
be
able
to
compose
your
own,
for
instance,
a
consistency
level
that
will
do
a
one
on
every
single
data
center
or
two
out
of
the
three
data
centers
or
a
specific
data
center
gets
quorum.
The
other
gets
one
very
different.
A
C
Protect
wise
is
basically
always
hiring
we're.
Looking.
We
love
tap
top
talent,
even
when
we
don't
have
a
wreck
open.
If
somebody
comes
along,
it's
really
awesome.
We
will
find
that
will
find
a
space
for
them
today.
We're
about
60
employees
we're
located
principally
in
Denver.
We
do.
We
do
lots
and
lots
of
big
data.
We
do
Network
processing
and
more
songs
real
time.
C
You
know
one
at
12
seconds
between
when
something
happens
on
your
network
and
you
get
a
finding
in
your
in
your
UI
and
we,
if
you
haven't
seen
our
UI
and
it's
worth
it's
quickly
time
to
go
check
it
out.
Rui
is
the
world
class
in
the
security
space,
we
believe
in
redesign
the
entire
user
experience
and
I
think
we've
done
a
really
excellent
job
of
that
we
have
some
of
the
best
node
engineers
around.
So
if
you
happen
to
be
a
node
guy
or
you
happen
to
be
a
deep
JavaScript,
id3
react
so
forth.
C
You
know
we'd
love
to
talk
to
you,
but
more
relevant
to
this
is.
We
are
always
looking
for
additional
people
in
the
Big
Data
space
in
the
stream
processing
space,
and
so,
if
you
have
have
or
even
are
interested
in
getting
into
big
data
stream
processing,
we
are
always
looking
for
bright
guys
in
that
space.
Awesome.