AWS Summit NYC 2012: Datapipe

Join Ed Laczynski, VP of Cloud Strategy at DataPipe, as he discusses deploying Turnkey Hybrid clouds with Amazon Web Services. DataPipe is a Mission-Critical IT Provider. Learn more: http://aws.amazon.com/solutions/case-...

---

Closed Caption:

skilled
vol 0
good afternoon everyone my name is Ed
licence key
I am from the fight to
have the privilege of being be the last
presentation today
saw Turkey was short and sweet hopefully
interesting and productive Tom little
bit about datapipe
we r a arm mission-critical IT services
provider
what does that mean means that we help
companies manage the most complex
IT infrastructure problems you come to
datapipe
when you have a complex problem
a multi-tier application or a mix of
cloud and
and non-party infrastructure you need
someone to manage a 24/7
you can sleep at night recognized in the
industry
as a leader in the space you be able see
that if the presentation
complete over itself but
you but pretty graphic on the right show
some love our
custom Rachael its we looked on it as a
sales pitch presentation but
are doing talk about some a bar
offerings may be relevant to the people
at this conference
om you know Amazon's particular summit
mixing
people in the community Amazon experts
and partners in the ecosystem in and
things are appropriate to talk about our
marriage Amazon program
I what we've done is we've taken our
managed service offering
which is everything up to and including
the application stack
so we go inside the OAS we go between
the OAS
we manage security we manage Tom
yes away and we've applied that to
Amazon infrastructure and we've had this
launch sense
think November 2010 has been very
successful program
Ross on interviews reseller so we can
bundle is all together into one package
for you
and were helping lot so far enterprises
adopt
cloud for this program taking over allow
the heavy lifting
from an Operations a national
perspective and also lost our apps as
well
that want to focus on development and
now on
car a restriction another service that
the proper to talk about that I don't go
into a lot more detail on
is our drug not enabled services and
that's where we're leveraging our
network
or leverage our capabilities to provide
I unique hybrid
it. solutions that that use
Amazon stricken active abilities I'll
talk about that
and more detail so what is higher in
I T well as hybrid om
I think a hybrid first thing about a
previous
or Nissan ball I companies on Baltimore
pure electric but something like that
right but not worth it with a hybrid and
and and
in the car business think I've arm a mix
a public cloud and anything else I am
whether it's on prime or the service
provider
anytime you're deploying something
wherever that unit
appointment is that relies on urs have
depended on her flows
flows data through I non
public our resources and
arm is 2012 so no longer
can I enterprise IT say
to the cool club guys get off my lawn we
don't want you here
with hybrid cloud I with hybrid IT
capabilities we can now
get the best of both worlds I'll talk
about that a little bit before going to
some the use cases I think
it supported to know what why wise is
available now why's this
wise hybrid IT hot all the sudden what
what's making this
relevant conversation I the number one
thing is that there's an ecosystem
has been built around Amazon and I a
proven provider skills
and products so you can deploy complex
solutions I
and reliable way arm from the enterprise
perspective
I public cloud has become a proven apart
other
above their strategy so there's a
confluence of events here
have allowed I whole industry built
around these are complex
IT problems second open access
networking
obviously arm without the ability to
connect your infrastructure
I into public cloud in a way that
letters of the low latency and
reliability in it's kinda hard core
level higher in
solutions so with the advent of Amazon
reckon acts
we now have this capability wouldn't
have in the past
and I think it's also important to note
how rawr us
has led in in coming out with a defacto
API standard so there is no industry
standard API for powered
right there's no consortium that's led
anything at this point so market-driven
and I was out like that and you know
it's appropriate here because look at
some like Amazon s3
for storage arm there's now the ability
to have
s3 like storage in your own facility
that
then backs up the amazon with one
communication mechanism the developers
to work on
and so s3 has become the standard I did
a lot that's driving
deductibility I've clout in general and
also
for hybrid scenarios
so what are some other challenges arm
that we can solve with higher I T I all
these are important
and I'll go through a couple use cases
for somebody's
but from things that way toward
compliance
a a lot of companies have very specific
revelatory regime than a need to
maintain
infrastructure our into an area that's
not yet public are ready for whatever
reason
and they need to build it also built a
career as a public cloud
activity to those resources when
appropriate
disaster recovery this is for a bleaker
on-ramp the cloud
if you're an enterprise scenario public
cloud
is a great place to start Dr Shirazi and
arm you know make sure you get that data
there
I in a reliable way with with the proper
our PO are to: arm is really important
and as available now
with things like I was in stores gateway
and service but as I did a pipe
and however managing research solutions
for customers
another be one is in Ohio
high-performance database solutions
I you know enterprise database solutions
are just not how compatible to tie
things at all
roxie enterprise so we'll talk about
something as well
we're seeing demand being driven I like
two different areas was a tactical area
and I was in for two drivers as well
so on a tactic side we're seeing
arm enterprises and
and a up you know organizations
that have pain in their on-premise or
hosted applications that
they need cloud to solids pass the bill
burst the cloud for example
arm any day a long-term archiving
solution
and they can afford to continue
investing
an cap that's heavy spinning disks arm
we're seeing folks at bill
their cloud infrastructure from the
ground up
in cloud also need some unique and
diverse
I hybrid solutions they may need to
deploy specific type I love
database or appliance or storage tiering
that
is not available out today and on
strategic side for sale idea center
outsourcing projects lot of folks that
are
basically saying I don't wanna run my
own data centers anymore I need to go to
cloud but I can't put a rain cloud
as fast as I need to make this move or
by the time I lease expires
so I need reliable activity any reliable
I am support for my entire portfolio
an I look at what I can move the cloud
and what I can move
to a up disarm a solution
anything to be connected and work
together so I went through
summers are a little more details on a
popular use cases were saying
and this year a higher high of storage
high-performance data bases be spoke
database services
with customers to come to us with a very
unique pre-existing
sir Big Ten architecture that can't be
changed very easily
and the cost them having a cloud
development team rebuild it
it is not worth it for them but they
want to take advantage of
up you know elasticity and on-demand
resources for us to the infrastructure
com off-site data backup coming from the
other way
we're seeing a lot of that customers
it's a I need to store
com my virtual machines by servers my
log files
I lost or the main cloud I want on
quickly reliably
and that's available to we'll talk about
that a little bit and
we're still seeing some custom hardware
integration like security
modules and Search Appliance -ism
also the weird stuff that goes on rocks
and doesn't quite fit
insider abiola youth report yet so that
kinda stuff is really drive a lot the
man
so the first scenario going a little
detail on is
offsite backup and Dom is rolling real
customer use cases that we have
arm with a customer that was on
up on physical infrastructure and they
wanna
the need to build back up there is a
global data somewhere and
and it was of was for redundancy and for
long-term archiving so
that a seven-year archiving requirements
and the cost again at maintaining
yes and power and all I kinda stuff
obviously doesn't make sense
anymore and issues part for om
one of the things that we did was
deployed in a way that
levers million see collectivity between
datapipe
facilities in Amazon's sauna scenario
stars so see the customers running
om Red Hat app servers
I in a facility there connecting
and Amazon be a drunken acts talk about
how that works
orbit and Dom that is being backed up in
store on Amazon
the Amazon search your and
to Amazon search keywords actually
presents I scars the lungs
honey technicians are here today but
you know to fairly standard wave and it
drives
and fairly straightforward a move data
so it's sir
enterprise friendly you going to
reengineer your architecture to
taken Angeles
another scenario is I'm high-performance
database scenario
and I believe this one is can talk about
Siebel Server Enterprise
om superstar enterprise requires a
quorum drive
anna & basically only kinda really works
well and and traditional infrastructure
scenarios and this customer was sorta
you know
baked into that solution and Dom
but wanted and a on-demand resources to
manage their billing cycles
so that a lot of did a process every
month
that in 10 pay for and manage Tom
physical infrastructure for those short
cycles an idiot to do it in a way that
lend itself to
reliability and low-latency connectivity
so in this case we arm
reengineer solution where we took their
oMG super enterprise environment we
collected into a single standard
environment and
I W us setup nearing which is
sauropods friendly over the wire moved
across and Lillian C
and then job provisioned ec2 instances
the process going to
skeezer this is running into millions it
is based transactions per day
as a large like social game actually the
game is
popular in South Korea but they run all
their stuff in the US
arm the
I the application leveraged
Security Group A W us we we made sure
that effectively was available between
the
the database in a data center and the
security group native to us
and we deployed in a way that the
and customer really you know they just
madison SL I that we provided and then
the up
their customer didn't see any difference
so they want when they want to the
billing server pick up the bill they had
no idea whether
they're coming from Amazon server or
server sitting
on prime it didn't really matter to them
and that's really key here you have to
do this and transparent way I think so
you know all the poor those things that
you can deploy successfully that you
know what
change your habits I think that's really
important so we need to do to deploy
this and I'm not sure
in the audience here is what IT managers
are or start up guys r or whoever
but Tom if you wanna you don't get into
deploy any sort of things
you know more than once even the first
time you need to
you know i think is on the key things
you should be thinking about om
get your getting expertise and or or fun
and expertise elsewhere
number one is II was under attack
capability said
Tom hybrid IT with Amazon is
is extremely powerful when you have that
network activity available
and we do see a tremendous amount and
demand for last customer see this as an
on-ramp the cloud
to save easy way to start and burst from
there
com doctor bowls from Amazon
I think or in the under French it heavy
lifting phrase and
arm for good or bad if you want to
deploy a hybrid IT
you need to be able to have those
capabilities so you need to be able to
manage or outsource our contract to
facilities
you have a supply chain for the bits and
pieces that are
gonna be assembled around the hybrid
components
inventory build capabilities et cetera
so
those things have a creep back into the
picture with hybrid om
for most enterprises they they probably
have some degree likability early today
dominant you know there's always service
providers in the ecosystem
I can supply it to you the networking
that's
are really oMG interesting and Tom
definitely not for the faint-hearted I
think
as as as we deployed our first couple I
Direct Connect based solutions hybrid
solutions we found that the networking
pease was
you know it different parts like
expertise
for go Janos is our never rounder and
you
runs our Direct Connect networking
services and it sa
the heat agree ahead it's not trivial so
have an ability to do never
configuration management a fairly
important charlie from MSP like us
we're doing hundreds of customers so the
problem said that we have is different
than a enterprise as one problem set so
thought impossible to something you
needed dive into those documents
and figure out and obviously you're
operating systems and software databases
in
all that stuff I the same stuff you have
to deal with in any environment
obviously need to consider it and make
sure your you're designing for
gonna SiC share for the
people who did show up today you're
seeing some secret
sauced of as much as I could get away
with sharing today
about how we've organize our director at
the table service network
with Amazon and very high level
obviously you can I going to be IP
addresses but it does show you how
up you know level complexity and also
that it's definitely doable and it's
real and it's available
I said before so in this diagram on the
left hand side you'll see
you see the Amazon network and things
like home
ec2 BBC I was ons
supplies appliances or services like
internet gateway
and a virtual gateway to unable activity
with the services on the right hand side
the sea
the five network in this case you know
both in a five and I was on our
provisioned through I he connects and
North Virginia and were geographically
close to each other
the on a dating website is the customer
solution I
the higher inside so it's gonna be your
database services restored services
you know there's any firewalls there in
all the same kinda stuff that we expect
to see
great side and then the middle part is
we're
interesting stuff is how you connected
things together so
om we use a technology called up you for
an over MPLS
ambulance pseudowire to allow for
be lands which are network domains
like a better word to trance and Amazon
and go into
customer solution or data you know if
you do this on your own will be your
view on anarchy is we have many of these
for each customer and it allows for
current activity between the two
services
although latency all manageable what
scale
again if you're and enterprise arm
IT manager if your startup because the
problems are solved
arm hopefully some this information is
useful for you
arm share with you help me go about RRR
central process for designing hybrid
are were it's very solution designed
heavy up front
is definitely a measure twice cut once
type arrangement
were the reasons why is just returned a
best practice it's nice to be able to
know what you're doing it for you start
doing it but also
on the hybrid side arm unlike ec2 I
don't have to throw server away
if something as a configured correctly
so we want to be sure of that
I and i know im so if you're if you're a
I said it. manager injured
offering this car service here
departments no def we won't have a
dialogue around solution design
we have some infrastructure design cover
configuration tools
excusing in-house that we use to to to
do that
com we've also built operating services
these provisioning services essentially
reserve
the space on Amazon side and on the
networks on
datapipe side for the services so
everything from
ensuring that we have governance over in
a US account to
understanding the provisioning
requirements for the network in
database for the storage all that's done
pre-built
am was built services and was built
services are essentially
actually implementing the work writes
about everything from configuring a
security group in a WS
were VPC to om
installing operating system or
application
we hand over solution run book again if
you're on the if you're on the inside
and your
offering us your customers especially
with cloud a great way to
social value i think is the hand back
over to your business owners I run book
or a description a solution so they
understand
the value provided also if there's an
emergency issues that sort of thing they
know what
the processes for remedying knows
and then we move into an operation
support mode numbers we operate this
hybrid IT infrastructure
common for
cell have it open up the questions
so have today I have about number
Coliseum did a pipe bomb
built on a BB business development Rob
Reiner senior
monarch attacked Kevin and I feel as
well so any questions or comments I have
an answer on these folks can as well
sure om you know you week we charge
you know the typical raws the Port Fairy
but only up up towards that oMG failure
preparing for so you may have a
now 210 good connections and you know
everywhere we can act and that's going
to be different pricing and
one single doing our shared connection
and then we have bomb
you know email standard pricing for
different service levels like databases
and servers at something talking happy
with our
but it's it's up acts just like a dove
you we try to blend it all together we
don't we'll
you know you have one bill Amazon data
points altogether you know
you not have to pick up access man you
know
use assume know what to do anymore
its hybrid so it's the you know it's
what we teach our achievements so
we have customers are running Oracle
rock and
in non Amazon connected into Amazon
and in all other app and web here's an
AW assets are managed under one
solution for pro yeah it's really a
question so
we com depends if you already had a US
account
week in we assume the account with a
prize a business process to do that and
then
Sorgen your bills from aus basically and
a neutral access our other services on
top of it
com we have a crandall storage on our
sin
I am for that purpose and we create the
account
if you're a new customer for example
will hand back you will hand you
you know provisioned keys and secrets
there a secure mechanism
from there as partners in process were
connected to the US and now
Virginia up the New Jersey which is more
long-term storage
and point and I San Jose and burn one
down as well
or in Asia as well we're not retractable
their
yet listed typical storage if you're
just looking for
com UBS like story was recommend he be
ass rate but if you're looking for
for tier 1 type stories that is
gonna be paid right into an Oracle rock
lobster some right that's completely
different
cost it is sold in a gigabyte a month
model just like it up your ass which I
wanna
sorry yes we do and we we issue BA's
you know it's a negotiated contract item
it's not a turnkey hit a button
everything be a IQ
shit scared