►
Description
Speaker: Aaron Stannard, CTO
The .NET ecosystem spent years on the sidelines, watching the NoSQL and distributed computing movements flourish in ecosystems like Java, Node.JS, and others.
Over the past year or so, the .NET ecosystem took matters into its own hands and has feverishly started adopting new ideas like NoSQL, reactive programming, the actor model, and more!
In this talk we're going to explore what the modern .NET enterprise stack looks like: Cassandra, Akka.NET, and Windows Azure. Also, we'll share what exciting new possibilities this has been able to create for some of the largest .NET shops in the world.
A
Good
morning
everybody,
so
my
name
is
Aaron
Stannard
I'm,
the
co-founder
and
CTO
of
peda
bridge
we're
a
company
that
helps
dotnet
enterprises
highly
available
distributed,
real-time
applications
and
I'm,
a
co-founder
of
the
aqua
dotnet
projects
and,
as
of
yesterday,
a
data
Stax
MVP
ray,
and
so
today
we're
gonna
talk
about
the
new
dotnet
stack,
and
so
this
talk
is
really
about
possibility
and
what's
available
to
us
in
the
dotnet
world
that
we've
never
had
before.
It's
not
really
gonna
be
a
low-level
like
architecture
talk
or
anything
like
that.
A
It's
gonna
be
about
at
a
high
level.
What
can
we
achieve
for
our
users
in
our
business,
with
some
of
the
new
options
and
tools
and
choices
that
are
available
to
us?
So
let
me
start
by
talking
about
the
experience
of
being
a.net
developer
and
what
it's
like.
So
this
is
share
a
little
bit
of
my
background
with
you.
You
know:
I
started
working
in
dotnet
in
2005
when
I
was
a
college
intern
working
for
fair
Isaac,
the
guys
to
do
your
FICO
score.
A
All
that
stuff
and
I've
been
in
dotnet
for
10
years,
I've
worked
at
Microsoft,
I've
started
three
companies
on
dotnet
and
they
founded
successful,
open
source
projects
on
it,
I'm
deeply
committed
to
this
ecosystem.
It's
the
reason
why
I
share
all
that?
Well,
one
of
the
things
that's
been
so
about
being
a.net
developer
is
we
have
to
play
by
this
guy's
rules
every
single
time
we
want
to
go
and
do
something
and
dotnet.
We
have
to
have
a
license
for
visual
studio
or
an
MSDN
subscription.
A
We
have
to
have
licenses
for
windows
for
our
development,
workstations
and
licenses
a
Windows
server
for
our
apps.
We
also
have
to
buy
sequel
spirit
licenses
and
be
really
precise
about
the
number
of
cores
and
the
types
of
features
were
gonna
need
when
we
develop
our
apps.
We
also
have
to
use
Microsoft's
web
stack,
asp.net
and
WCF,
because
everything
in
dotnet
runs
through
iis
and
windows
server,
and
so
the
ecosystem
has
been
a
place
where
you
can
really
sum
up.
A
The
experience
as
you
have
to
you
have
to
play
by
Microsoft's
rules
or
you
have
to
go
somewhere
else.
Let
me
share
a
little
store
with
you
briefly,
just
to
illustrate
the
reality
of
this.
So
when
I
was
working
at
Microsoft,
this
is
like
between
2010
and
2012.
In
early
2011,
html5
API
is
a
tall
band
fairly
standardized
and
browser
vendors
were
rapidly
trying
to
integrate
them
into
their
apps
and
the
most
exciting
one
that
got
developers
really
going
was
WebSockets,
because
they
enabled
a
whole
bunch
of
scenarios
that
we
could
have
never
done
before.
A
Being
able
to
have
a
true
server
push
inside
a
web
application
wasn't
something
we
really
had,
and
so
everyone
was
rushing
to
try
to
get
their
hands
on
WebSockets
as
fast
as
they
could.
But
dotnet
developers
had
a
little
problem
that
no
one
else
had,
which
is
that
is
seven
at
the
time,
did
not
support
WebSockets
IIAs
eight
would
but
that
wasn't
going
to
be
shipped
until
Windows.
Server
2012
is
every
dotnet
developer.
Who
wanted
to
try
to
use
WebSockets
to
deliver
a
better
experience,
their
end
users?
A
A
A
So
we
have
to
do
all
these
things,
or
at
least
we
did.
The
reality
is
that
in
the
past
couple
of
years,
dot
Nets
become
an
ecosystem
where
we
actually
have
freedom
and
choice
and
power,
and
we
can
go
and
really
deliver
experiences
beyond
what
we
were
limited
to
when
it
was
just
Microsoft's
world
Microsoft.
First
and
foremost,
you
heard
Scott
Guthrie
on
stage
yesterday.
They
fully
are
committed
to
the
idea
that
dotnet
has
to
be
bigger
than
them
in
order
for
it
to
have
a
future,
and
dotnet
can
no
longer
be
coupled
to
Windows.
A
So
one
of
the
things
that's
opening
up
for
us
is
the
possibility
or
we
could
build
highly
available
distributed.
Applications
in
pure
C,
sharp
and
F
sharp,
using
our
favorite
tools,
whether
it's
Visual,
Studio
or
maybe
even
sublime
text
on
OS
X.
We
can
go
into
populous
applications,
platform,
agnostic,
Lee
on
Windows
or
on
Linux,
and
we
can
use
the
most
modern
configuration
management
and
deployment
tools
to
do
it.
If
I
brought
up
this
possibility,
while
I
was
at
Microsoft
back
in
2012,
I
would
have
been
sent
to
a
charg.
A
A
A
With
the
advent
of
the
cloud
and
in
particular
Windows
Azure,
we
have
a
playground,
we
can
experiment
and
try
new
things
rapidly
at
minimal
cost.
You
saw
Scott
yesterday
fire
up
a
90,
no
DC
cluster
and,
like
three
clicks,
that's
amazing.
You
know
how
long
it
took
me
and
my
team
to
set
up
a
test
cluster
using
hyper-v
for
data
stacks
enterprise.
A
Honestly,
like
three
days
there
go
and
figure
out
all
the
different
little
things
we
had
to
do
to
go
and
make
it
work
and
editing
Yambol
files
and
all
the
rigmarole
of
manual
DevOps
with
Windows
Azure.
We
have
a
possible
future.
We
could
rapidly
try
and
test
stuff
and
learn
at
a
rate
that
is
unprecedented
for
software
developers.
A
Speaking
of
DSE,
the
next
part
of
our
stack
is
the
data
stack
and
one
of
the
things
that's
really
exciting
about
the
way.
Dotnet
is
right.
Now
is
every
single
major,
no
sequel
database,
whether
it's
Cassandra
or
MongoDB,
or
react
or
others.
The
old
tree
dotnet
like
big
kids,
now
we're
not
relegated
to
the
little
kids
table,
eating
peanut
butter
and
jelly
at
Thanksgiving
anymore.
We
can
sit
with
everybody
else
and
actually
enjoy
first
part.
Drivers
support
now.
A
So
that's
an
amazing
change
and
what
these
technologies
can
make
available
for
us
is
the
possibility
of
having
applications
that
can
perform
consistently
well
at
any
scale.
As
you
heard,
Jonathan
Ellis
mentioned
in
his
keynote
yesterday,
sequel
server
and
all
the
old
technologies
that
have
gotten
us
from
you
know
the
90s
until
now
are
built
on
a
model,
that's
fundamentally
brittle
and
that
models
not
going
to
keep
growing
with
us.
A
The
next
stack-
and
this
is
the
piece
that
has
me
most
excited
about
the
future
of
dotnet-
is
the
application
runtime,
so
quick
show
of
hands
here.
Who's
heard
of
core
CLR
raise
your
hands.
Okay,
about
quarter
of
you
core
CLR
is
an
initiative,
that's
being
developed
by
Microsoft
right
now,
that's
designed
to
be
cross-platform
and
docker
izybelle.
So
it's
a
version
of
the.net
runtime
that
supports
everything
you
need
to
deploy
in
asp.net
or
a
Windows
service
application
and
I
think
the
probably
had
support
for
WCF
and
other
stuff
too.
A
But
the
idea
is
that
we
can
actually
have
from
Microsoft
true
cross-platform
support
for
dotnet
and
really
make
it
a
multi-platform
runtime
and
the
other
thing
that's
going
on.
That's
really
exciting
is
what's
happening
with
mono,
since
unity3d
and
xamarin
have
become
giants
in
the
mobile
space,
monos
gotten
a
level
of
innovation
and
and
a
support
that's
never
had
before
and
with
Microsoft
open
sourcing.
The
CLR
recently
mono
4.0,
which
came
out
earlier
this
year,
actually
boasts
comparable
performance
to
the
CLR
under
a
number
of
use
cases.
A
Now
that's
never
been
the
case
in
Manos
decade-long
history.
It's
always
been.
You
know
the
slightly
laggy
cousin
of
the
CLR,
but
with
the
fact
that
Microsoft
is
opening
up
and
sharing
their
technology
they're,
making
the
experience
for
dotnet
on
platforms
that
aren't
Windows
better
across
the
board
and,
lastly,
well
actually,
no
I.
We
have
one
more
thing
after
that.
The
next
stack
here
is
our
application
stack
and
our
programming
model,
where
we
can
innovate,
show
of
hands
who's
heard
of
the
actor
model.
A
All
right.
Well,
that's
a
mate!
So
all
of
you,
dotnet
developers,
you
know
I'm
just
gonna,
go
ahead
and
keep
my
fantasy
here
and
assume.
Yes,
yes,
you
all
are
that's
amazing,
acha,
dotnet
and
akka
dot
cluster,
which
are
implementations
of
the
actor
model.
Allow
us
to
take
the
exact
same
sort
of
scalability
characteristics
that
cassandra
has
the
ability
to
be
always
up.
A
The
ability
to
add
new
hardware
on
demand
the
ability
to
recover
from
failures,
gracefully
akka
dotnet
gives
us
all
those
benefits
at
our
application
layer
and
with
it
and
its
programming
model
we
can
build
applications
that
would
have
been
considered
impossible.
A
couple
of
years
and
I'm
gonna
give
you
an
example
of
that
towards
the
end
of
this
talk
today
and
lastly,
the
front
end
so
asp.net
V
next
and
Nancy
effects
are
both
extremely
exciting
and
that
they're
cross-platform
runtimes
that
allow
us
to
use
the
best
tools
in
front-end
innovation.
A
So
a
good
example
working
on
a
Nancy,
FX
cross-platform
app
right
now
that
it's
able
to
use
Bower
and
type
scripts
and
able
to
use
a
bunch
of
JavaScript
libraries
and
frameworks
to
really
build
a
modern
application.
We
can
do
that
now
anywhere
in
our
front-end
stack.
We
don't
have
to
wait
for
Microsoft
to
go
and
add
a
little
feature
for
us
or
ship.
A
new
version
is
p.net.
The
state
of
the
art
is
already
such
that
we
can
use
all
those
tools
inside
our
app
and
build
the
best
possible
experiences
for
our
end
users.
A
So
the
future
is
bright.
With
this
new
stack,
we
have
the
opportunity
to
improve
and
innovate
and
really
create
something
new.
Every
single
area
in
our
technology
stack.
Let's
talk
a
little
bit
about
the
benefits.
Why
do
we
want
to
do
this?
Why
does
this
matter?
So?
The
first
benefit
is
that
we
get
to
break
free
from
the
sequel,
server
contention
pit
of
doom,
so
anyone
has
built
a
sufficiently
large
application
on
top
of
sequel.
A
You
end
up
in
this
situation,
where
you're,
giving
my
business
partner
Andrew
a
call,
saying:
hey,
yeah
we're
not
able
oozing
tens
of
thousands
of
dollars,
because
we
can't
keep
our
data
fresh
anymore
because
of
sequel,
server,
contention
and
timeouts.
One
of
the
things
this
stack
gives
us.
The
ability
to
do
is
break
free
from
that
model
where
there's
any
contention
to
begin
with,
or
we
can
go
and
use.
Potentially,
multiple
data
technologies
to
accomplish
different
things.
A
Jonathan
in
his
keynote
yesterday
mentioned
that
one
of
the
critical
things
we
have
to
do
as
modern
enterprises
is
segment.
Our
operational
database
away
from
our
analytical
reporting
database
that
those
are
orthogonal
concerns
and
one
of
the
things
that's
really
easy
to
fall
victim
to
and.net
is
blending
the
two
together
sequel,
server
reporting
services
is
the
ultimate
gateway
drug.
From
that
point
of
view,
so
one
of
the
things
that
we
can
do
with
the
new
dotnet
stack
is.
We
can
use
Cassandra
to
have
a
really
robust,
linearly
scalable
operational
database
for
supporting
our
apps.
A
That
can
give
our
users
a
consistently
fast
good
and
available
experience,
and
then
we
can
use
smart
streaming
to
actually
go
and
be
able
to
each
do
like
streaming,
ETL
workloads
and
gather
all
of
our
reporting
metrics
and
maybe
even
finally
write
them
back
into
sequel
server.
So
all
the
people
in
your
operations,
finance,
sales
and
marketing
departments
will
be
happy
to.
In
addition
to
your
users,
so
that's
something
that
opens
up
as
a
result.
A
A
Well,
this
is
a
future
that's
available
to
us
now.
If
we
can
really
get
a
hold
of
how
to
deploy
and
operate
our
software
on
linux,
using
technologies
like
docker
and
some
of
the
other
new
tools
that
Microsoft's
introducing
into
Windows
Azure
like
service
fabric,
we
can
start
to
do
some
really
cool
stuff.
A
The
next
piece,
the
tool
chain.
So
we
all
love,
Visual,
Studio
and
Microsoft
did
something
recently,
that's
really
remarkable,
which
is
they
made
Visual
Studio
Community
Edition
actually
include
all
the
features
that
pro
used
to
have
support
for
plugins
code
lens
and
a
lot
of
the
other
technical
tools
that
really
allow
us
to
be
super
productive
and
Visual.
Studio
that's
available
for
free
now
without
an
MSDN
license,
but
on
top
of
that
with
Visual
Studio
code,
the
new
browser
based
IDE,
they
announced
at
Build
conference
this
year.
A
We
can
actually
develop
our
apps
on
OSX
or
Linux
now
and
then
deploy
them
on
Windows.
If
we
want
and
then
who
can
forget,
omni
sharp
this
set
of
cross-platform
development
tools
that
Visual
Studio
code
uses
that
can
allow
us
to
go
and
develop
apps
with
full
and
tell
ascents
in
IDE
s
like
vim
or
sublime
text.
So
if
you've
got
someone
on
your
team
who's,
maybe
they're
a
recovering
Ruby
programmer
and
they
really
want
to
be
able
to
use.
You
know
all
their
little
Emacs
keyboard
shortcuts
and
everything
else.
A
You
know
how
those
those
IDE
those
text-
editor
people
are
right,
weirdos
anyway,
bottom
line
is
we
can
give
them
an
experience
where
dotnet
can
be
theirs
too.
So
we
have
choice
in
our
tool
chain,
but
then,
beyond
that
it
gets
even
better.
We
actually
have
open
source
frameworks
being
built
in
net
now,
like
akka,
dotnet
and
Helios,
and
dotnet
II
and
others
that
I
allow
you
to
do
things.
You
couldn't
do
before
like
distributed
application
development
like
really
good,
socket
programming
and
there's
more
and
more
that
are
coming
all
the
time.
A
As
a
result
of
some
of
the
changes
that
Microsoft
started
to
introduce
a
couple
years
ago,
the
dotnet
open
source
ecosystems,
starting
to
look
like
a
place
that
people
won't
make
fun
of
anymore,
because
if
there's
one
thing,
I
can't
stand
as
a.net
developer,
it's
being
pitied
by
JavaScript
developers,
so
our
tool
chain
is
becoming
exciting
and
speaking
of
which
did
I
mention
like
every
single
open
source.
You
know
development
framework
out.
There
is
starting
to
actually
try
to
support
net
data.
A
Stacks
is
an
amazing
driver
for
c-sharp,
solar
and
elasticsearch,
and
you
know
any
number
of
the
other
major,
no
sequel
technologies
like
MongoDB
and
react
all
support,
great
drivers
and
dotnet.
Now
you
don't
have
to
be.
You
don't
have
to
be
me.
Circa,
2013,
angrily
yelling
at
fluent
Cassandra
as
I'm
trying
to
figure
out
how
thrift
works.
A
So
we
have
an
availability
of
tools
that
we've
never
been
able
to
touch
before
outside
the
dotnet
ecosystem
and
that's
exciting.
We
are
totally
free
of
the
handcuffs
that
we've
had
when
the
when
the
show
was
totally
being
run
by
Microsoft's
product
release
cycle
that
excuse
of
you
having
to
wait
for
Microsoft.
That
weird
trade-off
would
have
to
make
we're.
Oh
we'll
catch
you
on
the
next
product
release
from
Redmond.
That's
no
longer
the
case
anymore.
A
We
can
be
agile
and
not
to
rip
on
JavaScript
developers
again,
but
it's
time
for
us
to
start
being
the
Predators
and
not
the
prey
anymore.
When
it
comes
to
innovation,
we
can
be
nimble,
we
can
be
fast,
we
can
experiment
at
a
scale.
That's
been
unprecedented
Annette,
so
there's
really
no
more
excuses.
A
Jason
Calacanis,
a
famous
investor
one
of
the
early
investors
in
uber,
a
number
of
other
companies
once
said
that
if
you're
not
building
a
five-star
app,
nobody
cares
and
that's
the
state
of
the
that's
really.
The
state
of
the
ecosystem
for
software
is
that
users
expect
us
to
be
available.
They
expect
us
to
be
responsive
and
they
expect
us
to
deliver
more,
and
we
have
to
do
that,
while
retaining
orders
of
magnitude
more
data
than
we've
ever
had
before.
A
With
this
stack,
we
actually
have
the
platform
to
go
and
do
that
at
a
scale
that
is
not
only
unprecedented
and
dotnet,
but,
as
I
mentioned
that
this
is
really
fun
to
develop
in.
This
is
an
opportunity
for
us
to
all
fall
in
love
with
dotnet
again.
So,
let's
dig
into
a
little
bit
more
detail
on
what
the
stacks
really
about
so
I
mentioned
that
the
stack
actually
begins
with
a
change
in
our
mindset
as
developers
so
in
education.
A
There's
these
two
ideas
of
a
fixed
mindset
and
a
growth
mindset,
and
let
me
tell
you
what
the
difference
between
those
two
is:
a
fixed
mindset
is
the
default
that
people
have
and
if
you,
if
you
have
children,
a
fixed
mindset
occurs
for
them.
When
you
praise
the
results
as
a
function
of
who
they
are,
you
know
if
I
did
really
well
in
a
spelling
bee.
A
My
mom
goes:
oh
you're,
a
genius,
you're,
great
speller,
then
I
believe
I'm,
a
genius
and
a
great
speller,
and
so,
if
I
get
put
into
a
situation
where
I
can't
spell
very
well
anymore,
it
threatens
my
identity
as
a
genius
and
so
I
go
and
retreat
back
to
my
comfort
zone.
Stay
in
that
little
fixed
world
view
that
I
have
about
myself.
Where,
as
a
growth
oriented
mindset,
is
you
praise
the
results
of
a
child
based
as
the
function
of
their
effort?
A
They're
gonna
come
out
of
it
with
all
the
effort
they
put
into
it,
and
so
how
that
applies
to
dotnet
is
partly
because
of
the
rails
that
Microsoft
has
had
us
on.
We
have
this
very
fixed
world
view
about
what
developing
for
dotnet
looks
like
and
what
sort
of
experiences
we
can
deliver.
Oh
well,
sequel,
server
and
asp.net,
you
know
crud
applications
can't
do
it
must
be
impossible
right
or
the
alternative.
Well,
if
we
can't
do
this,
you
know
with
this
setup.
A
Maybe
we
should
just
throw
a
lot
more
hardware
at
it
and
see
what
happens.
There's
a
sort
of
a
dearth
of
innovation
there,
because
we
really
haven't
had
a
lot
of
room
to
grow
and
actually
take
risks
and
experiments.
So
in
a
growth
oriented
mindset
we
can
stop.
Trying
to
you
know,
put
the
cart
before
the
horse
here
and
start
really
going
after
new
ways
of
building
applications
and
new
ways
of
structuring
our
teams
and
new
ways
of
deploying
our
experiences
and
delivering
them
to
customers
and
basically
put
everything
to
the
test.
A
So
growth
oriented
mindset
really
depends
on
having
choice
available
to
us
and
how
we
do
things,
and
we
have
that
now.
So
we
need
to
change
our
mindset,
so
we
can
take
advantage
of
it.
So
that's
where
the
platform
begins,
bringing
the
possibility
of
choice
and
and
improvement
to
your
team,
and
wouldn't
you
do
that
we
can
start
digging
into
the
real,
tangible
benefits
of
the
rest
of
the
stack
so,
for
instance,
with
Windows
Azure
Windows
Azure
offers
two
things
that
make
it
really
compelling
to
me.
A
My
last
startup
by
the
way,
was
built
totally
in
AWS,
so
I'm
a
hardcore,
Amazon
user,
so
I
know
what's
out
there,
but
what
Azure
has
that's
extremely
compelling
as
a
platform
is
their
new
service
fabric
platform
as
a
service
is
really
designed
to
build
enterprise
scale
applications?
Now,
as
Billy
said,
like
we're
not
all
trying
to
build,
you
know
cat
video
applications
out
there
right.
A
lot
of
us
are
trying
to
build
things
like
banking
infrastructure
or
being
able
to
do
Internet
of
Things
for
oil
and
gas
pipelines.
These
are
actual
use.
A
Cases
that
show
up
in
my
inbox,
at
beer
at
peda,
bridge
and
Microsoft
has
really
developed
azor
into
a
platform
where
you
can
support
those
sorts
of
use.
Cases
on
the
cloud
now
and
the
second
thing
that
got
going.
It
is
a
great
ecosystem
of
partners
like
data
snacks
and
10gen
and
Hortonworks
and
others
who
are
all
committed
to
being
able
to
deliver
open
source
experiences
for
dotnet
developers
on
Azure.
A
So,
with
that
ecosystem
in
mind,
we
can
treat
as
our
like
a
playground
to
rapidly
experiment
and
try
and
test
things,
and
even
if
you
know
if
your
business
still
has
its
own
data
center
and
it
needs
to
keep
some
things
on
premise
for
privacy
or
compliance
reasons,
you
can
still
use
Azure
for
the
parts
of
it
that
aren't
bound
by
that.
It's
not
all
or
nothing
game
right.
So
Azure
really
is
the
playground
that
we're
going
to
use
to
experiment
with
this
growth
mindset.
A
Next
is
Cassandra,
so
I
assume
you're
all
here,
because
you're
at
least
somewhat
aware
of
the
benefits
of
it
mean
you
know.
Hence
the
conference
right.
What
Cassandra
really
gives
us
that
we
never
had
with
sequel
server
is
a
scaling
model
and
a
reliability
and
availability
model.
That
makes
sense
and
is
easy
to
predict.
You
know:
I've
worked
with
I
have
a
DBA,
my
previous
job,
who
I
would
peg
him
in
the
top
500
sequence.
A
You
couldn't
you
couldn't
plot
that
on
a
histogram
for
you,
you
can
do
that
with
Cassandra
you're
gonna
get
a
linearly
scalable
performance,
which
means
that
your
load
relative
to
your
hardware,
if
you
manage
that
ratio
consistently
you're
to
deliver
consistently
fast
and
available
response
path
to
your
users
and
I'm,
going
to
give
you
some
metrics
at
the
very
end
of
this
talk,
you
can
bring
back
to
work
to
support
that
next
is
spark,
so
spark
might
be
the
most
exciting
thing.
That's
happened
to
the
data
ecosystem.
A
Hell
I
bet
half
the
sessions
going
on
at
Cassandra
summit
right
now
or
about
it.
What
spark
gives
us
the
ability
to
do
is
cleanly
separate
our
operational
metrics
from
our
analytical
ones.
We
can
go
and
keep
our
app
workload
on
Cassandra
and
have
consistent,
predictable
performance
there,
while
still
using
spark
to
deal
with
our
ETL
workload
and
our
analytic
workload,
and
we
can
write
it
back
out
to
tools
that
we
already
use.
A
God
knows
that
there's
a
massive
piece
of
infrastructure
built
on
top
a
sequel
server
for
reporting
and
analysis,
and
you
name
it
right
so
being
able
to
use
spark
to
allow
us
to
take
advantage
of
the
innovation
of
no
sequel
but
still
be
still
support.
Our
business
users
inside
our
own
company
is
an
amazing
asset
and
spark
has
similarly
linear
scalability
performance
too,
and
at
the
runtime.
So,
as
I
mentioned
before,
we
have
platform
choice.
A
This
is
the
thing
that
really
got
me
excited
about
being
a.net
developer
again
two
years
ago
was
that
I
could
build
applications
that
I
thought
were
impossible,
so
I,
the
real-time
marketing
automation,
platform,
I'm
gonna,
tell
you
about
and
I
think
three
slots
they
one
example:
we've
also
seen
people
building
multi
player,
games,
building
trading
platforms
and,
for
some
reason,
a
lot
of
them
build
gambling
systems
on
top
of
it.
I,
don't
know
why,
but
I
digress.
A
The
point
is
with
Akkad
on
net
and
the
programming
model
exposes
all
the
same
performance
and
scaling
characteristics
that
we
love
about.
Sequel
server
are
things
that
we
can
build
into
our
own
applications
now
and
lastly,
the
front
end:
a
quick
note
to
Microsoft
if
you
watch
the
recording
of
this
start,
adding
logos
to
your
projects,
if
you
want
them
included
in
talks
like
this,
so
Nancy
is
going
to
get
the
the
front
page
here.
A
The
bottom
line
with
our
front
end
stack
and
dotnet,
is
that
we
have
available
to
us
now
great
cross-platform
tools
for
doing
web
sockets
for
doing
web
api
and
doing
this
simple
web
applications.
Nancy
and
asp.net
v-necks
and
signal
are
and
web
api
all
work
equally
well
on
linux,
as
they
do
on
windows,
and
you
should
see
some
of
the
metrics
microsoft's
putting
out
about
kestrel
the
new
web
server
that
hires
asp.net
v.
Next,
it
might
outperform
iis
and
no
short
period
of
time.
So
there's
a
lot
of
exciting
innovation
happening
here.
A
For
me
at
my
previous
company
marked
up,
so
we
were
a
real-time
analytics
and
marketing
automation,
platform
for
Windows,
Store,
Windows,
Phone
and
Windows
desktop
developers,
so
that
embed
our
SDK
into
their
app
and
would
allow
them
to
do
things
like
send
targeted
offers
to
users
based
on
what
their
behavior
was,
and
one
of
the
things
that
we
learned
that
we
had
to
be
able
to
do
with
our
marketing
automation
product
was.
We
had
to
be
able
to
serve
those
offers
within
seconds
like
three
seconds
of
a
user
qualifying
for
a
campaign.
A
Otherwise
the
performance
of
that
campaign
would
decrease
by
an
order
of
magnitude.
Like
that's.
The
difference
like
30
seconds
was
the
difference
between
a
four
percent
conversion
rate
versus
three
seconds,
which
would
be
a
forty
percent
conversion
rate.
We
were
able
to
measure
that
so
we
picked
this
stack,
the
new
Donette
stack,
we
used
Amazon,
Web,
Services,
Cassandra,
Akkad
net
and
asp.net,
and
here's
what
were
able
to
deliver,
30
million
distinct
devices
per
month
would
connect
to
our
service
and
they
would
write
upwards
of
500
Meg's
of
event.
A
Data
per
second,
these
aren't
images
going
into
a
CDN.
This
is
the
actual
stuff
hitting
our
load
balancer
for
our
REST
API
and
going
through
our
application
at
peak
hours.
That
is
a
lot
of
information,
and
that
resulted
in
about
a
hundred
million
database
transactions
per
day,
and
that
includes
batch
right.
So
we're
modifying
you
know
30
or
40
counters
at
once,
and
you
know
what
we
could
handle
that
entire
workload
on
30
like
medium
sized
servers
on
Amazon.
A
You
know,
run
concurrently
across
multiple
machines,
so
it
has
this
amazing
ability
to
a
demise
and
reduce
the
size
of
your
code.
So
a
quick
question
who
thinks
that
code,
that's
83%
smaller,
is
going
to
be
at
least
83%
easier
to
debug
how
about
83%
easier
to
train
your
developers
on
how
about
you
know.
The
list
goes
on.
A
There's
enormous
benefits
to
being
able
to
do
this,
the
turnaround
time
to
build
this
entire
marketing
automation
product
from
an
idea
on
a
sheet
of
paper
in
the
back
of
a
napkin
to
actually
being
in
production
with
six
months.
You
know
the
team
of
three
developers,
but
the
amount
of
fun
we
had
developing
on
the
stack
was
amazing.
A
I
think
is
that
ten
thousand
percent
per
hundred
thousand
more
and
it's
a
it's-
it's
really
difficult
to
express
just
how
different
the
experience
of
being
a
Donna
developer
felt
when
we
were
able
to
use
Cassandra
we're
able
to
use
a
DSC
search
and
solar,
we're
able
to
use
Aqua,
Net,
and
now
that
you
know
this
is
a
couple
of
years
ago
that
we
did
this
the
day.
I
could
go
and
run
this
on
Linux
with
docker
and
I
could
go
in
use.
You
know,
asp.net
be
next
to
go
run.
A
You
know
all
the
all
the
fun
front-end
stuff
that
we
had
to
try
to
do
like
cobble
together
by
hand
before
this
snack
is
even
better
now
than
it
was
a
couple
years
ago.
But
these
are
some
pretty
incredible
numbers.
I've
shown
you
so
I
want
to
end
today's
talk
with
a
question.
What's
next
for
you
and
what
could
be
made
possible
for
you?
If
you
brought
this
to
your
company
and
made
it
work
and
I
would
assert.
A
A
B
So
yeah
you
mentioned
mano
as
part
of
this,
and
it's
always
been
that
little
bastard
stepchild.
So
can
you
elaborate
a
little
bit
more
on
your
saying
so
they're
really
gonna
bring
that
up
or
is
the
common
clr
I
mean
cuz.
It
seems
like
there's
always
gonna,
be
that
contention
between
the
two
right
yeah.
A
That's
a
great
question,
so
what's
Manos
role
in
a
world
where
core
CLR
is
being
developed
and
here's
my
assertion
on
that
the
core
CLR
is
still
gonna
be
released
at
the
speed
of
Microsoft,
even
though
the
core
CLR
is
being
developed
out
in
the
open
and
it's
open
you're
still
going
to
be
subject
to
the
yeah.
Microsoft
is
three
thousand
lawyers.
Man,
they're
they're
gonna
travel
at
a
speed.
That's
not
even
comparable
to
mono
the
role
that
I
think
mono
has.
A
Is
it's
trying
to
be
more
like
the
full
footprint
of
dotnet,
but
on
POSIX
based
systems?
So
mono
should
be
able
to
support
anything
that
you
can
do
in
the
CLR
on
POSIX,
whereas
the
core
CLR
it
is
designed
to
be
about
ten
percent
the
size
of
net?
It's
a
really
small
footprint,
so
it's
like
forty
mags
versus
four
hundred
and
that
so
it
could
be
docker
eyes
and
deployed
multiple
times
on
the
same
host
machine
right.
A
So
the
different
rule
difference
between
the
two
I
think
is
gonna
come
down
to
which
parts
of
the
framework
do
you
need
if
you're
just
running
a
web
application?
And
you
don't
need
anything
crazy,
like
on
a
C++,
C
alive
or
being
able
to
do
P
invoke
or
anything
like
that,
then
pour
CLR
will
get
the
job
done,
but
if
you've
got
a
little
heavier
duty
application-
and
maybe
it
has
some
legacy
stuff
built
into
it-
mono
is
gonna,
be
the
tool
of
choice
for
that
in
Linux
and
mono
is
definitely
not
going
away.
A
Unity,
3d
and
xamarin
are
too
big
to
let
let
anything
happen
to
it.
So
I
think
when
you're
gonna
continue
to
see
is
that
mono
is
always
gonna
be
available
and
it's
always
gonna
be
innovated
on,
and
it's
really
up
to
really
up
to
the
community
so
to
speak,
to
say:
well
what
do
they
want
to
do
with
it
on
the
server
and
overwhelmingly?
There's
a
lot
of
push
from
companies
that
Andrew
and
I
work
with
to
actually
try
to
move
off
of
Windows.
B
A
Current
support
bracket
on
net
and
core
CLR,
so
we
have
one
contributor
in
a
fork:
who's
actually
got
most
of
it
running
I.
Think
the
thing
is
trying
to
work
on
now.
Is
the
networking
stack,
so
we
have
all
the
core
actor
framework
running
on
Quora
CLR,
but
the
stuff
like
Akkad,
a
persistence,
aa
Kedavra
mo
a
cannot
cluster
anything
that
really
depends
on
either
a
database
driver
or
a
Helios.
A
The
socket
server,
all
that
stuff
needs
to
be
ported
a
core
CLR
in
order
for
it
to
work,
so
we
don't
officially
support
core
CLR
yet,
but
that
is
going
to
change
as
core
CLR
matures,
because
it's
still
in
an
alpha
state
right
now
and
our
goal.
You
know
this
to
give.
You
are
sort
of
State
of
the
Union
Anaka
net.
We
think
core
CLR
should
be
the
default
weighed
and
applied.
Server-Side
developers
build
their
applications.
A
We
think
that's
gonna
be
the
case
over
the
next
18
months,
but
we
need
to
see
progress
in
core
CLR
first
before
we
start
moving
the
framer
that
way,
because
where
most
of
our
users
are
right
now
I
mean
you
know
me
often
I
get
asked
about
dotnet
3,
5,
&,
4
Oh
support
for
Akkad
on
net,
it's
like
daily,
so
we've
got
a
little
bit
of
work
to
do
on
bringing
the
rest
of
the
ecosystem.
Out
of
you
know,
2008
before
we
before
we
jump
ahead
to
2016
yeah.
A
A
Are
some
pain
points
using
Microsoft
Azure?
Do
we
have
time
for
another
minute,
talk
the
answer?
Some
of
the
pain
points
cloud
services
from
Liz
Loza,
the
current
version
of
Azure
pass
for
anything.
That's
not
a
stateless
web
application
using
sequel.
Azure
is
a
living
hell
to
work
with,
and
the
reason
why
I
say
that
is
one
of
the
things
you
should
never
do.
If
you're
building
developer
tools
is
try
to
be
really
smart
about
what
the
other
developers
trying
to
do,
and
so
azure
will
go
and
do
things
like.
A
Oh
we'll,
just
go
ahead
and
start
rebooting
servers
and
all
that
sort
of
stuff
and
update
the
hypervisor
under
the
hood,
and
so,
if
you're
like
me
and
I,
do
stateful
application
development
doing
that.
With
this
unpredictable,
you
know
chaos
monkey
rolling
around
nuking
my
servers
every
month.
It
can
be
really
difficult.
On
top
of
that,
there's
a
lot
of
little
management
stuff.
They
do
under
the
hood.
A
That
doesn't
make
full
sense
to
me
and
then
the
last,
the
most
painful
part
with
cloud
services
is
the
fact
that
I
have
to
basically
take
my
visual
studio
solution
and
totally
rewrite
it
in
order
to
work
with
Windows
Azure
at
all,
I've
got
to
go
and
wrap
all
my
windows
services
into
a
worker
role
project
and
do
all
that
stuff.
So
the
experience
around
that
is
still
kind
of
kind
of
gnarly
with
service
fabric.
A
The
new
version
of
past
the
releasing
you'll
be
able
to
deploy
like
docker
eyes
like
docker
images
and
run
with
those
which
will
make
things
a
lot
nicer
than
it
is
now.
So
that's
so
that's
one
pain
point
that
I
think
is
addressed
already
and
the
things
Microsoft's
announced,
but
I've
personally
and
actually
pretty
recently
been
pretty
frustrated
about
that
particular
one
all
right.