[Full slides now also available at http://www.slideshare.net/adrianco/netflix-on-cloud-combined-slides-for-dev-and-ops]
Short summary of why Netflix is running on the Amazon cloud, what is running there, what we have learned and where this is taking us.
This is the introduction section to a series of public presentations that will go into much more detail. The Silicon Valley Cloud Computing Meetup was on Oct 14th, QCon San Francisco November 3rd.
1. Ne#lix
in
the
Cloud
Oct
14,
2010
Adrian
Cockcro:
@adrianco
#ne#lixcloud
h=p://www.linkedin.com/in/adriancockcro:
2. Beta
Slides
"I
didn't
have
-me
to
write
a
short
le4er,
so
I
wrote
a
long
one
instead.”
Mark
Twain
Teaser
intro
slides
now
at
slideshare.net/adrianco
Oct
14th
Beta
#devops
subset
-‐
Cloud
CompuKng
Meetup
Nov
3rd
GA
–
QConSF
and
full
slides
on
slideshare.net
3. With
more
than
15
million
subscribers
in
the
United
States
and
Canada,
Ne#lix,
Inc.
is
the
world’s
leading
Internet
subscripKon
service
for
enjoying
movies
and
TV
shows.
.
4. Synopsis
• Why
Give
This
Talk?
Why
Use
AWS?
• The
Goals
– Faster,
Scalable,
Available
and
ProducKve
• Datacenter
AnK-‐pa=erns
– the
things
we
wanted
to
change
and
why
• Cloud
Architecture
Features
• Cloud
Bring-‐up
Strategy
• Developer
TransiKons
and
Tools
• Roadmap
and
Next
Steps
5. Why
Give
This
Talk?
• Ne#lix
is
Pathfinding
– Cloud
ecosystem
is
evolving
very
fast
– Share
with
and
learn
from
the
cloud
community
• We
want
to
use
clouds,
not
build
them
– Cloud
technology
should
be
a
commodity
– Public
cloud
and
open
source
for
agility
and
scale
• We
are
looking
for
talent
to
help…
– How
do
we
connect
with
the
very
best
engineers?
– h=p://www.quora.com/10X-‐Engineers
6. Why
Use
AWS?
• We
stopped
building
our
own
datacenters
– Capacity
growth
rate
is
acceleraKng,
unpredictable
– Product
launch
spikes
-‐
iPhone,
Wii,
PS3,
XBox
– Datacenter
is
large
inflexible
capital
commitment
• Leverage
AWS
Scale
–
the
biggest
public
cloud
– AWS
investment
in
tooling
and
automaKon
– AWS
zones
for
high
availability,
scalability
– AWS
skills
are
common
on
resumes…
• Leverage
AWS
Feature
Set
–
most
advanced
– EC2,
S3,
SDB,
SQS,
EBS,
EMR,
ELB,
ASG,
RDB..
7. “The
cloud
lets
its
users
focus
on
delivering
differen-a-ng
business
value
instead
of
was-ng
valuable
resources
on
the
undifferen)ated
heavy
li0ing
that
makes
up
most
of
IT
infrastructure.”
-‐
Werner
Vogels
August
25,
2009,
‘All
Things
Digital’
8. Ne#lix
Deployed
on
AWS
• “batch”
Movie
Encoding
farm
(2009)
– Thousands
of
EC2
instances,
Petabytes
of
S3
– Movie
files
staged
out
to
CDNs
for
delivery
• Hadoop
-‐
ElasKc
Map-‐Reduce
(2009)
– Large
scale
log
processing
and
analyKcs
• Streaming
Service
Back-‐end
(early
2010)
– Highly
available
and
scalable
“play
bu=on”
• Web
site,
a
page
at
a
Kme
(through
2010)
– Ramp
from
25%
of
views
to
80%
in
the
coming
weeks
• API
for
TV
devices
and
iPhone
etc.
(2010)
– Personalized
movie
choosing
algorithm
back-‐end
9. Learnings…
• Datacenter
oriented
tools
don’t
work
– Ephemeral
instances,
high
rate
of
change
• Cloud
Tools
Don’t
Scale
for
Enterprise
– Built
our
own
tools,
drove
vendors
hard
• “fork-‐li:ed”
apps
don’t
work
well
– Fragile,
too
many
datacenter
oriented
assumpKons
baked
in
• It’s
faster
in
the
end
to
re-‐code
than
Knker
– Re-‐architected
and
re-‐wrote
much
of
our
code
base
– Fine
grain
web
services
– Leveraging
open
source
in
Java
– SystemaKcally
instrumented
– “NoSQL”
SimpleDB
backend.
In
the
datacenter,
robust
code
is
best
prac-ce.
In
the
cloud,
it’s
essen-al.
10. Next
Few
Years…
• “System
of
Record”
moves
to
Cloud
– Cut
the
datacenter
to
cloud
replicaKon
link
• InternaKonal
Expansion
–
Global
Clouds
– Rapid
deployments
to
new
markets
• GPU
Clouds
opKmized
for
video
encoding
• Cloud
StandardizaKon
– Cloud
features
and
APIs
should
be
a
commodity
not
a
differenKator
– DifferenKate
on
scale
and
quality
of
service
– CompeKKon
drives
cost
down
– Higher
resilience
– Higher
scalability
We
would
prefer
to
be
an
insignificant
customer
in
a
giant
cloud
11. Takeaway
NeIlix
is
path-‐finding
the
use
of
public
AWS
cloud
to
replace
in-‐house
IT
for
non-‐trivial
applica-ons
with
hundreds
of
developers
and
thousands
of
systems.
h=p://www.linkedin.com/in/adriancockcro:
@adrianco
#ne#lixcloud