5. AWS Global Infrastructure
10 Regions
consisting of
26 Availability Zones
and
52 Edge Locations (CDN)
Customer Decides Where Applications and Data Reside
6. AWS Region View
- Independent/Separate Geographic Areas
- Isolated from other Regions (security boundary)
- = ~50 mile radius “clustered” data center architecture
- Comprised of multiple Availability Zones
- Availability Zone = 1 or more “data center”
- Availability Zones connected through redundant low-latency
links
- Customer chooses a Region and Data stays within Region.
- Enables High-Availability Architecture
Sample US Region
Availability
Zone A
Availability
Zone B
Availability
Zone C
7. AWS Availability Zone (AZ) View
- Multiple Isolated locations within a Region
- Availability Zone = 1 or more “data center”
- Independent Failure Zone
- Physically separated
- On separate Low Risk Flood Plains
- Discrete UPS
- Onsite backup generation facilities
- Fed from different segments of utility provider
- Redundantly connected to multiple tier-1 ISP’s
- No “Disaster Recovery Datacenter”
- Built for Continuous Availability
- Customer decides Availability Zone for Compute
Sample US Region
Availability
Zone A
Availability
Zone B
Availability
Zone C
~ Data Center
12. DR & business continuity
DR is part of a wider set of policies and controls…
High availability Backup Disaster recovery
It’s not an all or nothing thing
Choose what needs to failover and what does not
Some things more important than others
Some things will still be working
Keep your applications
running 24x7
Make sure your data is protected
and can be recovered if it is lost
Get your applications and
data back after a major
disaster
13. DR & business continuity
Each set of IT assets will have different requirements…
Recovery Time Objective
(RTO)
How quickly you need this asset to be recovered?
e.g. 1min? 15min? 1hr? 4hrs? 1day?
Recovery Point Objective
(RPO)
How ‘fresh’ the recovery must be for the asset?
e.g. zero data loss, 15mins out of date?
14. DR & business continuity
Assets will sit on a spectrum of technical complexity…
Rebuild when
required from
offsite backup
Run hot-hot
configuration with
auto-failover
16. Traditional, second datacenter
The fundamental economic model…
Primary Site
Routers
Firewalls
Network
Application Licenses
Operating Systems
Hypervisor
Servers
SAN fabric
Primary Storage
Backup
Archive
Secondary Site
Routers
Firewalls
Network
Application Licenses
Operating Systems
Hypervisor
Servers
SAN fabric
Primary Storage
Backup
Archive
17. Utility, on-demand datacenter
The fundamental economic model…
Primary Site
Routers
Firewalls
Network
Application Licenses
Operating Systems
Hypervisor
Servers
SAN fabric
Primary Storage
Backup
Archive
AWS
Routers
Firewalls
Network
Application Licenses
Operating Systems
Hypervisor
Servers
SAN fabric
Snapshot Storage
Backup
Archive
Secondary
site costs
18. Business & technical drivers
With utility services you might be able to:
Reduce costs
Slash DR budgets by up to 50%
Reduce on-premise
Eliminate 30%+ of on-premise
physical equipment
Consolidate sites
Eliminate the need to run a
secondary site
Remove aging
technologies
Eliminate tape for backup and
archive
19. Conventional DR Sites
Challenges around Cost
High Cost
Low ROI
Implemented only for
most critical systems
Usually scaled down to
50% of production
Systems in a remote
region challenging
20. Disaster Recovery on AWS
Cost Effective – On Demand Infrastructure
Unprecedented
capabilities to implement
DR sites
Easily set up DR sites on
different geographic
regions
Cut down DR site cost by
up to 70%
Substantial savings on
software licenses
22. Amazon
Simple
Storage
Service (S3)
AWS Import/Export
AWS Storage
Gateway Service
AWS Direct
Connect
Amazon Virtual
Private Cloud
(VPC)
Amazon
Route 53
Amazon Elastic
Compute Cloud
(EC2)
Amazon Relational
Database Service (RDS)
Amazon
Elastic Block
Storage (EBS)
Object storage &
transfer services
Networking services Foundation services
23. AWS storage is ideal for DR
S3 and Elastic Block Store
Simple Storage Service
Highly scalable object storage
1 byte to 5TB in size
99.999999999% durability
Elastic Block Store
High performance block storage device
Volumes of 1GB to 1TB in size
Mount as drives to instances with
snapshot/cloning functionalities
24. Durable
Designed for 99.999999999%
durability of archives
Glacier
Cost effective
Write-once, read-never. Cost effective for long
term storage. Pay for accessing data
3 to 5 hour Retrieval time
25. Direct Connect
Dedicated connection between your IT
infrastructure and the AWS datacenters
Extend your network infrastructure and
VLANs into AWS
VPN Connection
A Hardware VPN connection connects
amazon environment to your datacenter
Internet Protocol security (IPsec) VPN
connection
Commonly used hardware supported
Internet
Virtual Private Cloud
Private, isolated section of the AWS Cloud
Launch resources in a virtual network that you
define
complete control over your virtual networking
environment
Internet
Connecting to AWS
27. Common DR architectures
4 main patterns
Backup & Restore Pilot light
Warm standby in
AWS
Multi-site solution in
AWS & on-premise
28. Common DR architectures
Let’s start with Backup & Restore
Backup & Restore Pilot light
Warm standby in
AWS
Multi-site solution in
AWS & on-premise
29. Backup & Restore pattern
Advantages to starting a journey with this pattern
Simple to get started
Easy starting point for exploring the AWS cloud
Low technical barrier to entry
Focus on incorporating cloud into your DR
strategy, not on complex technical issues
related to hot-hot systems
Cost effective
Very high levels of data durability at low price
Cost of storing snapshots in S3
Archiving possibilities beyond tape using Glacier
30. Backup & Restore pattern
The preparation process…
Take backups of
current systems
Store backups
in S3
Move to long term
archive in Glacier
31. Backup & Restore pattern
The process…
Take backups of
current systems
Store backups
in S3
Detail how you will restoring from backup or
recover from archive
Move to long term
archive in Glacier
34. Common DR architectures
Let’s look at the Pilot Light pattern…
Backup & Restore Pilot light
Warm standby in
AWS
Multi-site solution in
AWS & on-premise
35. Pilot light architecture
Moving along the DR spectrum…
Build resources
around replicated
dataset
Keep ‘pilot light’ on by replicating core
databases
Build AWS resources around dataset
and leave in stopped state
Scale resources in AWS
in response to a DR
event
Start up pool of resources in AWS when
events dictate
Match current production capacity
through auto-scaling policies
38. Common DR architectures
Let’s look at the Warm standby pattern…
Backup & Restore Pilot light
Warm standby in
AWS
Multi-site solution in
AWS & on-premise
39. Warm standby architecture
Moving along the DR spectrum…
Build resources
around replicated
environment
Operate a warm standby by replicating
app servers and core databases
Build AWS resources around dataset
and run in limited capacity
40. Warm standby architecture
Moving along the DR spectrum…
Build resources
around replicated
environment
Operate a warm standby by replicating
app servers and core databases
Build AWS resources around dataset
and run in limited capacity
Scale resources in AWS
in response to a DR
event
Scale up pool of resources in AWS when
events dictate
Match current production capacity
through auto-scaling policies
43. Common DR architectures
Let’s look at the Multi-site pattern…
Backup & Restore Pilot light
Warm standby in
AWS
Multi-site solution in
AWS & on-premise
44. Multi-site architecture
Moving along the DR spectrum…
Deploy resources
necessary to operate
full production
Operate a full stack by replicating app
servers and core databases
Fail over to AWS in
response to a DR event
Sufficient resources in AWS to handle full
peak load
48. Resources
Disaster Recovery on AWS: aws.amazon.com/disaster-recovery
Architecture Center: aws.amazon.com/architecture
Using AWS for Disaster Recovery
http://media.amazonwebservices.com/AWS_Disaster_Recovery.pdf
Backup and Recovery Approaches Using AWS
http://media.amazonwebservices.com/AWS_Backup_Recovery.pdf