►
Description
Speaker: Amit Khare, NoSQL Architect
Company: BetterCloud
BetterCloud works on Insights, Security and Management for cloud office platforms like Google Apps and Office 365. In order to support large data sets with no impact to data retrieval and data ingestion performance, we are using Cassandra as the Data Storage engine to serve our Customers across the globe.
A
Alright,
so
great
background,
so
they're
louder
around
about
three
green
hangers.
We
let
this
be
the
number
one
Google
Apps
provider
for
her
company
going
for
email
and
collaboration
with
games.
Us
left
that
company
started
dinner
club
real
estate
services
needed
product
numbers
on
west
market
places.
We,
since
we
launched
our
product
and
our
partition,
has
expanded
when
it
counted
just
focusing
on
google
apps
we're
also
looking
at
office
365,
because
any
company
that's
going
to
be
go
to
the
cloud
toward
email
collaboration.
You
believe
that
those
are
going
to
be
wanted
to
provide
authorities.
A
A
Fifty
fifty
thousand
companies
are
smaller
products
representing
30
million
users
and
a
lot
of
countries,
and
the
reason
we
have
so
much
data
is
for
every
one
of
those
users
were
going
proven.
Getting
the
metadata
for
dots,
there's
Google
Sites
their
calendar
secondaries
mount
with
Google+
posts.
Third
party
apps.
They
can't
install
it's
just
a.
B
Beep
spotted
off
at
the
App
Engine
storage,
that's
where
we
start
their
birth.
At
this
point
of
time,
300
we
are
set
up,
0,
90
degrees
of
data
who's
that
that's
not
really
most
efficient
at
all
for
us,
so
we
we
started
researching,
some
of
which,
which
can
be
cost
effective
for
us
and
that's
when
I
drove
to
Oakland
sociopathy
Cassandra
as
we
started
our
journey,
the
family
skeptical
of
moving
on
to
the
central.
We
definitely
want
a
linear
scalability
without
a
Jordan
accountability
of
ours.
B
Since
senator
standard
about
Cassandra
there's
a
certain
issue
carnivals
where
there
were
certain
colleges
that
priests
are
increasing,
demeaned,
Italian
is
most
rapidly
developing
each
and
every
person
cause
of
the
water
fire.
Cassandra
is
and
getting
them
up
to
speed
go
to
the
documentation
which
is
right
by
datastax.
That's
excellent,
I
we're
going
to
do
all
the
wheel.
You
have
the
sessions
with
developers,
never
get
the
good.
B
So
that's
where
we
are
right
now,
I,
we
are
in
personality,
good
Cassandra
and
work
under
challenge
that
we
started
facing
was
being
from
win
from
of
the
items
are
gone
under
the
alaska
said.
How
should
we
warn
of
the
realization
that
we
have
I?
Should
we
retrieve
them
successfully
with
all
that
without
so,
and
that
one
of
that,
whenever
those
without
any
knowledge
of
charge,
Sandra
dishonored,
creating
in
Texas
directly
on
the
tables
on
which
the
keys
will
not
visit?
But
both
local
area
performance
started,
started
coming
down?
B
My
data
set
from
the
other
transaction
fee,
but
then
that
became
more
complex
from
the
public
is
not
strategy
standpoint,
and
then
we
started
looking
for
better
options
who
secondary
indexes
was
again
an
option
which,
which
were
using
definitely
at
this
one
of
nine
but
cheaper
certain
cases
like
four
million
fields,
which
is
it
indefinitely
it
before
measure
when
it
goes
on
to
the
good
in
fields.
But
you
only
change
y,
which
was
not
a
different
exit
for
the
food
it
feels
was.
We
are
not
storing
the
coupon.
B
B
It
reduces
the
indexing
cpu
usage
because
the
center
does
not
relate
next
knowledgeable.
We
could
happen.
Just
indexes
the
fog
balance
and
the
third
part
we
we
went
ahead
and
that's
where
the
need
arise
for
the
end
up
going
and
indexing
I.
We
wanted
each
and
every
little
peach
apricot
inflated
apart,
and
we
wanted
some
some
questions
involving
the
nearest
exit
president's
own
ethnic
provider.
For
a
few
more
realistic
too.
B
That's
where
we
started
dyslexic
man,
certain
certain
distributions
like
without
those
will
strike
you
as
regular
Cassandra,
so
studying
the
central
coordinate
deserts.
It
basically
likes
all
the
index
list
on
to
the
new
scenes.
Debussy
and
Exeter
forced
those
indexes
into
the
same
force
pocket
center,
so
it
I
read
in
slowly.
B
So
the
right
one
more
than
this
way
is
able
to
you
get
a
write-up
in
Cassandra.
It
flows
and
I
dramatic
is
not.
They
were
stored.
It
you
get
using
the
custom
checkin
in
Texas
part.
The
second
will
next
limousine
Renee's
elusive
miles
and
death.
When
retrieval
mechanism,
it
first
pulls
off
the
new
season,
texted
and
said
who's
walked
across
fornicator.
So
that's
how
that's
how
it
works
and
the
best
part
of
that
was.
B
There
was
one
single
right
path
and
there
was
particularly
fun
biosecurity,
so
you
can
basically
read
all
the
gotham's
print
along
polygons
that
you
want
to
directly
from
the
sakura
by
having
the
news
in
search.
Queries
as
part
of
the
sequencer
fan
will
never
worry
about
anything
else.
You
only
visit
challenge
with
spread
yo
Cassandra
was.
B
B
So
we
went
ahead
and
we
saw
that
there
are
no
solution
to
it.
It's
got
a
Stargate
metaphysical.
What
that
does
is
it
does
a
similar
battle.
What
strategy
does,
but
it
maintains
different
lab
reason,
please
Jacques
of
its
own,
and
you
can
just
use
those
java.
That's
part
of
the
lovely
food.
Is
it
commander
and
use
your
own
custom
in
texas
to
begin
creating
files
of
renton
pretty
fabulous,
so
not
not
cooperative,
and
we
are
very
happy
with
it
for
four
minutes
of
some
of
the
performance?
B
Does
it
cost
house
was
pretty
good,
but
when
it
comes
to
when
it
comes
with
a
distributed
nature
of
of
the
large
amount
of
paint
job,
we
definitely
want
some
such
depth,
which
will
boast
with
which
were
boosting
in
texas
separately
and
the
readers
operating
on
to
the
center.
So
Megan
example
just
like
three
decentralized,
as
opposed
the
sooner
it
takes
us
the
preview
of
the
solute,
and
it
was
a
data
or
so
that's.
B
Of
concept:
oh,
why?
Having
cleared
up
in
house,
guess
the
condensing
so
that
whenever
you
were
quite
upset,
it
should
I
Neely's
right
directly
onto
the
mechanism
like
Kafka
or
a
vernacular
finish
on.
It
is
excused
or
rub
it
directly
to
Nationals.
So
that's
what
we
started
researching
upon
act
prevented
all
those
ingredients
or
whatever
we
spoke
with
at
the
documentation
on
the
festivities.
Now
again
that
and
this
one
of
time
where
we
are
pretty
much
go
further,
I
think
we
would
do
it
for
sure
at
this
one
of
them.
B
We
are
already
nominated
grinding
back
to
Dhaka
directly
from
his
hat.
Also
write
code
word.
If
not
its
you
will
know,
you
will
not
need
to
write
full
flags
not
to
when
Cassandra
went
to
kalka.
Just
a
record
is
a
man
would
be
basically
push
it
off
with
a
flag
that
you
can
use.
Any
good
news
is
not
good
as
soon
as
those
partners
growth
anyway.
So
that
was
very
good.
B
So
I
bet
well,
we
tend
to
do
the
test
for
another
print
and
it's
basically
a
current
address.
Ppd
and
the
integration
dresses
players
are
the
data
sources
there
so
that
whenever
the
data
model
changes,
you
must
ensure
that
all
the
changes
that
are
taking
it
can
part
of
the
data
model
on
the
integration
this
poop
and
then
you
can
go
through.
So
it
really
helps
us
to
effectively
at
learning
timorous,
frequently
and
forward
without
an
it
is
issues
that
we
get
faced
on
the
team.
B
It's
a
good
thing,
but
sure
what
one
thing
that
that
we
ought
we
as
well
are
awarded,
who
was
being
on
my
internet
application,
for
you
also
want
to
prac
each
and
every
planet
little
bit.
For
example.
What
is
the?
What
is
a
number
of
schemes
that
is
there
for
trading
forex,
great
customer
organic?
Then
what
is
the
usage
of
the
application
onto
them?
B
B
Those
things
and
the
last
thing
that
that
the
challenge
we
face
of
the
sea
shops
and
what
was
the
propeller,
might
shake
mix
that
we
use
it
for
jakarta
connections.
It
returns,
actually
an
object
and
does
not
really
return
an
interface.
So
it's
not
very
helpful
for
us.
So
Ossetia
happened
who's
sitting
out
here,
yes,
yellow
metal.
Where
he's
going
to
focus,
focus.
B
B
So
here's
an
example
of
basically
the
connector
is
basically
increase,
support,
sorry,
Kenneth,
multi-touch,
but
sure
that
we
have
it
was
pinky
to
be.
We
use
the
clustering
tears
because,
on
your
ring,
difference,
Cleo
Nicholas
Ricky,
when
you
always
wanted
the
default
order
to
be
part
of
the
time
plus
lately
and
texture
than
another,
actually
enforce
the
concatenation
of
strength
of
news
rain
and
it
is
baby,
so
it
automatically
sorts
all
the
data
sets
or
to
the
biz
rhenium
focus
in
ascending
order
and
then
uses
dog
get
ready
for
the
which
makes
it
unique.