With the increased use of cloud services, organizations are faced with finding the most efficient way to use existing IT infrastructure alongside cloud-based compute, storage and networking resources. This has resulted in the rise of hybrid IT whereby companies leverage both on-premises and cloud resources to drive increased agility, stability and accessibility.
Ask The Architect: RightScale & AWS Dive Deep into Hybrid IT
1. Ask
the
Architect
Diving
Deep
into
Hybrid
IT
Brian
Adler,
Sr.
Services
Architect,
RightScale
Miles
Ward,
Senior
Manager,
SoluAons
Architecture,
AWS
Claudio
GenAle,
Cloud
SoluAons
Engineer,
RightScale
5. Four
Tracks
for
Business
Value
Increase Enterprise Agility
Drive Operational Efficiencies
Expand Markets & Capabilities
Increase Investment Flexibility
Faster Time to Market
Accelerate Innovation
Fail (or Succeed) Fast
Cut Infrastructure Costs
Reduce Overprovisioning
Increase IT Efficiency
New Geographies
New Capabilities
Web-Scale Applications
CapEX to OpEX
Reduce sunk costs
Reduce lock-in
7. RightScale’s
View
of
Cloud
Management
A
framework
of
tools
to
solve
cloud
challenges
and
increase
your
ability
to
enjoy
cloud
benefits
!
• Consistency
and
reliability
of
shared,
commodity
resources
• Secure
access
to
a
variety
of
remote,
networked
resources
• Self-service
accessibility
of
an
approved
configuraAon
menu
• ElasAc,
instantaneous
scalability
driven
by
automation!
• Usage
transparency
and
optimization!
8. The
Core
Cloud
Management
Toolset
Resource
management
• Compute,
network,
storage
Governance
• Policy
management
• IdenAty
and
access
management
• Usage
and
cost
management
Configura/on
• ConfiguraAon
management
• Service
catalog
• License
management
Automa/on
• Resource
monitoring
and
alerAng
• Automated
scaling
and
deployment
orchestraAon
10. EvaluaAng
ApplicaAons
for
the
Cloud
Business
impact
• Should
we
put
this
app
on
the
cloud?
• What
benefits
will
we
get?
Technical
requirements
• Can
we
put
this
app
on
the
cloud?
• How
much
will
migraAon
or
build
out
cost?
Choose
the
quick
wins!
• High
business
impact
• Appropriate
technical
requirements
• Start
greenfield
and
worry
about
migraAon
later
14. Use Cases
Cloudburs/ng
• An
applicaAon
deployment
model
in
which
an
applicaAon
runs
in
your
on-‐premises
environment
and
bursts
to
a
cloud
provider
when
the
demand
for
compuAng
capacity
spikes.
• The
advantage
of
a
cloud
bursAng-‐enabled
deployment
is
that
an
organizaAon
only
pays
for
extra
compute
resources
when
they
are
needed.
Considera/ons:
• Control
–
public
Internet
traversal
may
be
required
• Latency
–
traversal
of
public
Internet
involves
the
Great
Unknown
• Cost
–
bandwidth
charges
for
public
Internet
traversal
• Complexity
–
numerous
opAons
exist
for
configuraAon
of
a
hybrid
secure
environment
15. Hybrid
IT
/
CloudbursAng
ON-PREMISES
CLOUD
LOAD BALANCERS
APP SERVERS
MASTER DATABASE
SLAVE DATABASE
OBJECT STORAGE
APP SERVERS
PUBLIC
INTERNET
CloudbursAng
16. Use
Cases
Disaster
Recovery
(DR)
• ProducAon
environment
in
on-‐premises
environment
• DR
environment
in
the
cloud
• Most
common
configuraAon
is
the
“Warm
DR”
scenario
• ReplicaAng
slave
in
the
cloud
• All
other
servers
in
non-‐operaAonal
state
18. Use
Cases
Self-‐Service
IT
Portal
(“IT
Vending
Machine”)
• Users
select
one
of
several
pre-‐configured
tech
stacks
• Isolated
Test/Dev
environments
20. On-‐Premises
IT:
Key
ConsideraAons
Cost
• Sunk
investment
in
physical
infrastructure
–
try
to
maximize
the
uAlizaAon
of
this
CAPEX
Workload
and
Infrastructure
Interac/on
• ApplicaAons
have
different
resource
needs
• Choose
the
right
fit
for
your
applicaAon
and
your
infrastructure
Compliance
• Data
may
be
contained
on-‐premises
for
compliance
needs
or
requirements.
21. On-‐Premises
IT:
Key
ConsideraAons
Latency
• Consumers
of
the
on-‐premises
resources
are
generally
“closer”
to
the
environment,
which
reduces
latency
User
Experience
• Related
to
latency,
end
user
experience
is
enhanced
due
to
proximity
to
resources.
22. Management
ConsideraAons
Elevates:
• Configurations
• APIs
• Automation behaviors
• Access control
• Billing and governance
Users and systems
RightScale – Multi Infrastructure Resource Management
Common
API
and
UI
AWS
On-Premises OS
Hardware Hypervisor Services
APIs
APIs
23. Design Considerations
Location of Physical Hardware
• On-premises
• Availability considerations (power, cooling, networking, etc.)
• Hosted or Co-location facility
• Accessibility of hardware for additions and/or modification
• Latency to end users
• Security
24. Design
ConsideraAons
Availability and Redundancy Configuration
• Easiest configuration (single zone, single region, single API endpoint) does not
promote high availability
• High Availability of resources requires more complex configurations
26. Design
ConsideraAons
Intended
Workloads
and
Use
Cases
• Does
the
applicaAon
require
high
availability
or
is
it
tolerant
of
interrupAons
of
service?
• Does
the
applicaAon
require
(or
greatly
benefit
from)
specialized
hardware
or
network
configuraAons?
31. Customer
Example
#1:
HA/DR
Issue
faced:
• Highly-‐available
web
applicaAon
architecture.
Disaster
Recovery
environment
with
RPO
as
low
as
possible,
and
RTO
of
under
two
hours.
SoluAon
developed:
• MulA-‐AZ
applicaAon
with
redundant
components
in
two
or
more
AZs
per
Aer.
DR
environment
in
separate
AWS
region.
Architecture
• Three-‐Aer
(four,
with
caching)
in
AWS
US
East
• “Warm
DR”/”Pilot
Light”
in
separate
AWS
region
32. Customer
Example
#2:
CloudbursAng
Issue
faced:
• Private
cloud
with
limited
capacity.
Web
applicaAon
with
unpredictable
demand,
and
prone
to
viral
events.
SoluAon
developed:
• On-‐premises
environment
“bursts”
into
AWS
US
East
VPC.
Architecture
• Three-‐Aer
applicaAon
with
scalable
array
in
AWS,
and
secondary
slave.
• Direct
Connect
configuraAon
between
on-‐premises
and
AWS
US
East
33. Customer
Example
#3:
IT
Vending
Machine
• Issue
faced:
• Sales
Engineers
need
to
be
able
to
provision
mulA-‐node
deployments
for
Hadoop,
allowing
customers
to
evaluate
the
soluAon
for
a
specified
period
of
Ame.
Deployments
need
to
be
automaAcally
cleaned
aier
that
Ame.
• SoluAon
developed:
• Web
UI
consuming
the
RightScale
API
to
assemble
required
deployments,
tracking
and
cleaning
them
aier
the
evaluaAon
period.
• Architecture
• N/A
34. User identity Your portal
On-Premises
end users
(web browser)
request a certain server type is
launched, and specify cloud location
Amazon Web Services
Deploy the chosen server
in a private or public cloud,
based on the
ServerTemplate
Service catalog of IaaS
applications
(ServerTemplates)
RightScale API
RightScale UI
service
designers(web
browser)
server
instance
Self
Service
IT
Portal
server
instance
is displayed as service
catalogue items within
review access logs, version
control and change logs, set
permissions
managers (web
browser)
billing data sent to
RightScale, for all cloud
usage
36. Q&A
• [I
am]
looking
for
reusable
templates,
best
pracAces,
and
approaches
to
perform
assessments
on
moving
workloads
to
the
cloud.
• RightScale
MulACloud
Marketplace
• Repository
of
hundreds
of
publicly-‐available
ServerTemplates
• Onboarding/CloudSight/CloudRank
37. Q&A
• How
does
applicaAon
management
work
in
hybrid
clouds
where
the
private
cloud
workload
may
burst
to
AWS?
What
are
good
pracAces
for
app
data
sync
when
the
app
is
running
in
private
cloud
and
AWS?
• Stateless
applicaAon
servers
are
best-‐suited
to
cloudbursAng.
• If
data
synchronizaAon
is
required:
• Highly-‐available
shared
storage
(such
as
S3)
• High-‐performance
shared
filesystem
(like
GlusterFS
or
OpenAFS)
• Implement
a
low-‐latency
secure
link
(such
as
DirectConnect).
38. Q&A
• Is
there
a
published
checklist
somewhere
of
what
all
needs
to
be
validated
prior
to
deciding
to
move
a
system
(Exchange,
AD,
etc.)
to
the
cloud?
And
is
there
a
checklist
for
what
all
needs
to
take
place
when
you
actually
decide
to
move
something?
Do
you
have
any
whitepapers
to
show
how
customers
have
been
impacted
posiAvely
by
moving
on
prem
systems
to
cloud
that
most
customers
have
(i.e.
Exchange,
Email,
etc.)
How
have
they
saved
money,
provided
beler
performance,
etc.),
and
made
their
lives
easier?
• This
is
a
common
quesAon
we
get
at
RightScale
-‐-‐
so
common
that
we
developed
a
consulAng
product
offering
around
for
this
exact
need
(CloudSight).
This
is
tough
to
answer
in
general
terms,
but
would
be
a
great
topic
for
our
in-‐person
events.
39. Q&A
• How
does
using
Amazon
VPC
impact
using
right
scale
middle
ware?
• Instances
need
to
route
out
to
RightScale
(no
inbound
access
required).
• NAT
• EIPs
• Route
through
on-‐premise
device
40. Q&A
• How
to
deal
with
long
distance
latency
issues
in
data
transfer?
• WAN
OpAmizaAon
Tools
• Compression/De-‐dup/TCP
OpAmizaAons,
etc.
41. Q&A
• If
we
save
data
to
the
AWS
cloud,
does
the
key
reside
on
the
cloud
or
with
us
exclusively?
(for
security
reasons,
we
would
prefer
not
to
have
the
key
reside
on
the
cloud
as
it
does
on
some
systems)
42. Upcoming
“Ask
the
Architect”
Events
OperaAonalize
Your
AWS
Usage
Through
An
Enterprise
IT
Vending
Machine
Webinar
AWS
re:Invent
Online
9/17
Connecting your Datacenter to AWS
Designing and Implementing a Hybrid IT
Strategy and Solution
@
AWS
re:Invent
Las
Vegas,
NV
11/12
44. Agenda
10:30AM
-‐
11:00AM
RegistraAon
11:00AM
–
11:15AM
IntroducAons
11:15AM
–
12:00PM
Planning
for
a
Hybrid
IT
Environment
12:00PM
–
01:00PM
Lunch/Networking
01:00PM
–
02:00PM
Planning
for
a
Hybrid
IT
Environment
(conAnued)
02:00PM
–
03:00PM
Q&A
45. Hybrid
IT
Trends
Source: RightScale State of the Cloud 2013
Currently running apps in
public cloud
Currently running apps in
on-premises environments
Currently running apps in
hybrid environments
Current Cloud Usage
By organizations with a Hybrid IT strategy
Enterprises want multi-cloud/hybrid IT
Hybrid IT
47%
Multi-private
15%
Multi-public
15%
46. DefiniAons
and
Terminology
On-‐Premises
IT
A
collecAon
of
compute,
storage,
and
network
resources
for
a
single
tenant
that
are
accessed
programmaAcally
via
an
API
endpoint.
Cloud
Architecture
A
similar
set
of
resources
that
is
mulA-‐tenant
and
is
provided
by
a
cloud
vendor
with
access
via
an
API
endpoint.
Hybrid
IT
An
environment
that
spans
your
cloud
provider
as
well
as
one
or
more
on-‐premises
IT
environments.
Mul/-‐Cloud
An
environment
that
spans
two
or
more
separate
cloud
providers
or
regions
48. Enterprises
are
AdopAng
AWS
to
Achieve
the
Benefits
of
the
Cloud
Replace
CapEx with
OpEx
✔
Lower
Overall
Costs
✔
No More
Guessing
Capacity
✔
Agility /
Speed /
Innovation
✔ ✔
Shift Focus to
Differentiation
Go Global
in Minutes
✔
49. Many
Enterprises
Worry
that
These
Are
the
Only
Two
Choices
Build a new
‘private’
cloud.
Rip everything
out and move
to AWS.
#1 #2
50. The
Good
News
Is
that
the
Cloud
Isn’t
an
‘All
or
Nothing’
Choice
Corporate
Data
Centers
On-premises
Resources
Cloud
Resources
Integration
51. Active Directory
Network Configuration
Encryption
Back-up Appliances
Users & Access Rules (IAM)
Your Private Network (VPC)
Encryption (S3, RDS, HSM)
Back-ups (Storage Gateway)
AWS Direct
Connect
Your On-premise Apps Your Cloud Apps
IntegraAng
AWS
with
Your
ExisAng
On-‐premises
Infrastructure
Corporate
Data
Centers
52. The
AWS
Ecosystem
Allows
You
to
Use
Your
ExisAng
Management
Tools
Single Pane of Glass
On-
premises
Data
Centers
Consulting &
Management Partner
54. Strategy
1:
Cloud
for
Development
and
Test
Environments
SAP
Reduced
deployment time
from weeks to days
SAP
70% reduction in
operational costs
Oracle
Reduced dev &
test environment
costs
“Using cloud for Test
and Development
workloads is an easy
win. A lot of enterprises
are using AWS in this
way first.”
James Governor,
Founder
55. Strategy
2:
Build
New
Apps
for
the
Cloud
Faster to build
Facebook
app
Global
web sites
Mobile
streaming
Social
games
Consumer
apps
Genetic
sequencing
Marketing
campaigns
Less expensive to run
Distributed
architectures for high
availability
Easier to manage
Financial
record
archiving
57. Operational
applications
Export data to
AWS for analytics
processing
Oracle
databases
Automated back-
up to S3 with
Oracle RMAN
Strategy
3:
Use
Cloud
to
Make
ExisAng
On-‐Prem
Apps
Beler
Retail
operational
applications
Data
synchronization &
back-up with AWS
storage gateway
58. Strategy
4:
New
Cloud
Apps
that
Integrate
Back
to
On-‐Prem
Systems
On-
premises
Data
Centers
Hybrid
App
AWS serves up
application content &
data
Integration back to
Samsung Data
Centers for financial
transactions
59. Strategy
5:
Migrate
ExisAng
Apps
to
the
Cloud
1/3 of servers
migrated to AWS
Saved 1-1.5 million
GBP
App
Your Data
Center
Migrated 500 web
properties in 5
months
New product web
sites in 2 days vs. 2
weeks
Migrated clinical trials
simulations platform
Simulations in 1.2hrs vs.
60hrs
64% reduction in costs
App
60. Strategy
6:
All-‐In
10,000s of EC2 instances in
multiple regions & zones
100s of middle tier services &
applications to support 33M+ global
members
At peak consumes 1/3 of US
Internet bandwidth
69. On-‐Premises
IT:
Key
ConsideraAons
Cost
• Sunk
investment
in
physical
infrastructure
–
try
to
maximize
the
uAlizaAon
of
this
CAPEX
Workload
and
Infrastructure
Interac/on
• ApplicaAons
have
different
resource
needs
• Choose
the
right
fit
for
your
applicaAon
and
your
infrastructure
Compliance
• Data
may
be
contained
on-‐premises
for
compliance
needs
or
requirements.
70. On-‐Premises
IT:
Key
ConsideraAons
Latency
• Consumers
of
the
on-‐premises
resources
are
generally
“closer”
to
the
environment,
which
reduces
latency
User
Experience
• Related
to
latency,
end
user
experience
is
enhanced
due
to
proximity
to
resources.
71. Management
ConsideraAons
Elevates:
• Configurations
• APIs
• Automation behaviors
• Access control
• Billing and governance
Users and systems
RightScale – Multi Infrastructure Resource Management
Common
API
and
UI
AWS
On-Premises OS
Hardware Hypervisor Services
APIs
APIs
72. Design Considerations
Location of Physical Hardware
• On-premises
• Availability considerations (power, cooling, networking, etc.)
• Hosted or Co-location facility
• Accessibility of hardware for additions and/or modification
• Latency to end users
• Security
73. Design
ConsideraAons
Availability and Redundancy Configuration
• Easiest configuration (single zone, single region, single API endpoint)
does not promote high availability
• High Availability of resources requires more complex configurations
75. Design
ConsideraAons
Intended
Workloads
and
Use
Cases
• Does
the
applicaAon
require
high
availability
or
is
it
tolerant
of
interrupAons
of
service?
• Does
the
applicaAon
require
(or
greatly
benefit
from)
specialized
hardware
or
network
configuraAons?
76. Use Cases
Cloudburs/ng
• An
applicaAon
deployment
model
in
which
an
applicaAon
runs
in
your
on-‐
premises
environment
and
bursts
to
a
cloud
provider
when
the
demand
for
compuAng
capacity
spikes.
• The
advantage
of
a
cloud
bursAng-‐enabled
deployment
is
that
an
organizaAon
only
pays
for
extra
compute
resources
when
they
are
needed.
• ConsideraAons:
• Control
–
public
Internet
traversal
may
be
required
• Latency
–
traversal
of
public
Internet
involves
the
Great
Unknown
• Cost
–
bandwidth
charges
for
public
Internet
traversal
• Complexity
–
numerous
opAons
exist
for
configuraAon
of
a
hybrid
secure
environment
77. Hybrid
IT
/
CloudbursAng
ON-PREMISES
CLOUD
LOAD BALANCERS
APP SERVERS
MASTER DATABASE
SLAVE DATABASE
OBJECT STORAGE
APP SERVERS
PUBLIC
INTERNET
CloudbursAng
78. Use
Cases
Disaster
Recovery
(DR)
• ProducAon
environment
in
on-‐premises
environment
• DR
environment
in
the
cloud
• Most
common
configuraAon
is
the
“Warm
DR”
scenario
• ReplicaAng
slave
in
the
cloud
• All
other
servers
in
non-‐operaAonal
state
80. Use
Cases
Self-‐Service
IT
Portal
(“IT
Vending
Machine”)
• Users
select
one
of
several
pre-‐configured
tech
stacks
• Isolated
Test/Dev
environments
81. User identity Your portal
On-Premises
end users
(web browser)
request a certain server type is
launched, and specify cloud location
Amazon Web Services
Deploy the chosen server
in a private or public cloud,
based on the
ServerTemplate
Service catalog of IaaS
applications
(ServerTemplates)
RightScale API
RightScale UI
service
designers(web
browser)
server
instance
Self
Service
IT
Portal
server
instance
is displayed as service
catalogue items within
review access logs, version
control and change logs, set
permissions
managers (web
browser)
billing data sent to
RightScale, for all cloud
usage
82. The
Business
Case
for
Moving
to
the
Cloud
IT Needs The AWS + RightScale Solution
The ability to scale your project portfolio
without the need to scale headcount.
• Professional services to design and implement.
• Provision predictable and repeatable AWS and on-
premises resources.
• Ensure compliance on all IT resources.
Ensuring 24x7 uptime in a global
marketplace.
• Advanced monitoring and alerting.
• Automated remedial actions across all your IT
resources.
Maintain visibility and control. Advanced reporting and user access providing visibility
and control across all your IT resources.
Consolidated management of
infrastructure environments.
A single pane of glass for managing both all your IT
resources.
83. Upcoming
“Ask
the
Architect”
Events
OperaAonalize
Your
AWS
Usage
Through
An
Enterprise
IT
Vending
Machine
Depth
of
content
Webinar
Digging Deep with Hybrid IT
AWS
re:Invent
In
Person
Online
9/17
Los
Angeles
9/10
NYC
10/1
Boston
10/15
Connecting your Datacenter to AWS
Designing and Implementing a Hybrid IT
Strategy and Solution
@
AWS
re:Invent
Las
Vegas,
NV
11/12
84. Q&A
• [I
am]
looking
for
reusable
templates,
best
pracAces,
and
approaches
to
perform
assessments
on
moving
workloads
to
the
cloud.
• RightScale
MulACloud
Marketplace
• Repository
of
hundreds
of
publicly-‐available
ServerTemplates
• Onboarding/CloudSight/CloudRank
85. Q&A
• How
does
applicaAon
management
work
in
hybrid
clouds
where
the
private
cloud
workload
may
burst
to
AWS?
What
are
good
pracAces
for
app
data
sync
when
the
app
is
running
in
private
cloud
and
AWS?
• Stateless
applicaAon
servers
are
best-‐suited
to
cloudbursAng.
• If
data
synchronizaAon
is
required:
• Highly-‐available
shared
storage
(such
as
S3)
• High-‐performance
shared
filesystem
(like
GlusterFS
or
OpenAFS)
• Implement
a
low-‐latency
secure
link
(such
as
DirectConnect).
86. Q&A
• Is
there
a
published
checklist
somewhere
of
what
all
needs
to
be
validated
prior
to
deciding
to
move
a
system
(Exchange,
AD,
etc.)
to
the
cloud?
And
is
there
a
checklist
for
what
all
needs
to
take
place
when
you
actually
decide
to
move
something?
Do
you
have
any
whitepapers
to
show
how
customers
have
been
impacted
posiAvely
by
moving
on
prem
systems
to
cloud
that
most
customers
have
(i.e.
Exchange,
Email,
etc.)
How
have
they
saved
money,
provided
beler
performance,
etc.),
and
made
their
lives
easier?
• This
is
a
common
quesAon
we
get
at
RightScale
-‐-‐
so
common
that
we
developed
a
consulAng
product
offering
around
for
this
exact
need
(CloudSight).
This
is
tough
to
answer
in
general
terms,
but
would
be
a
great
topic
for
our
in-‐
person
events.
87. Q&A
• How
does
using
Amazon
VPC
impact
using
right
scale
middle
ware?
• Instances
need
to
route
out
to
RightScale
(no
inbound
access
required).
• NAT
• EIPs
• Route
through
on-‐premise
device
88. Q&A
• How
to
deal
with
long
distance
latency
issues
in
data
transfer?
• WAN
OpAmizaAon
Tools
• Compression/De-‐dup/TCP
OpAmizaAons,
etc.
89. Q&A
• If
we
save
data
to
the
AWS
cloud,
does
the
key
reside
on
the
cloud
or
with
us
exclusively?
(for
security
reasons,
we
would
prefer
not
to
have
the
key
reside
on
the
cloud
as
it
does
on
some
systems)
• TrendMicro,
a
RightScale
partner,
has
SecureCloud
which
allows
data
encrypAon
with
the
key
management
provided
by
Trend’s
SaaS
platorm
(not
in
AWS),
or
in
your
own
datacenter.
90. The
Business
Case
for
Moving
to
the
Cloud
IT Needs The AWS + RightScale Solution
The ability to scale your project portfolio
without the need to scale headcount.
• Professional services to design and implement.
• Provision predictable and repeatable AWS and on-
premises resources.
• Ensure compliance on all IT resources.
Ensuring 24x7 uptime in a global
marketplace.
• Advanced monitoring and alerting.
• Automated remedial actions across all your IT
resources.
Maintain visibility and control. Advanced reporting and user access providing visibility
and control across all your IT resources.
Consolidated management of
infrastructure environments.
A single pane of glass for managing both all your IT
resources.