0 best practices for architecting for the cloud 1. Enable Scalability 2. Use Disposable Resources 3. Automate Your Environment 4. Loosely Couple Your Components 5. Design Services, Not Servers 6. Choose the Right Database Solutions 7. Avoid Single Points of Failure 8. Optimize for Cost 9. Use Caching 10. Secure Your Infrastructure Everywhere Speaker: Anson Shen

1. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Eric Lam
Solutions Architect – AWS
June 2018
Architecting for The Cloud
- Best Practices

2. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #1: Enable Scalability
Ensure that your architecture can
handle changes in demand.

3. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Become Stateless
• Don’t store state in server
• Leverage services to hold
state information
• Application functions
regardless of which
application node processes
the request

4. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
#2: Use Disposable Resources

5. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #3: Use Disposable Resources
Think of servers and other components
as temporary resources.
Job completed?
Shut the
instance down

6. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Availability Zone 2
RDS DB
instancestandby
(multi-AZ)
RDS DB
instance
Availability Zone 1
Elastic Load
Balancing
Elastic Load
Balancing
Auto-Scaled
EC2
Instances
Auto-Scaled
EC2
Instances
AMI, Tools &
Scripts

7. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Think Parallel
Hour 1 Hour 2 Hour 3 Hour 4
• One Server working for Four
hours costs the same as Four
servers working for One hour
• Combine with elasticity to
increase capacity when you
need it most
• The beauty of the cloud
shines when you combine
elasticity and parallelization

8. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Availability Zone 2
RDS DB
instancestandby
(multi-AZ)
RDS DB
instance
Availability Zone 1
Elastic Load
Balancing
Elastic Load
Balancing
Auto-Scaled
EC2
Instances
Auto-Scaled
EC2
Instances

9. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Availability Zone 2
RDS DB
instancestandby
(multi-AZ)
RDS DB
instance
Availability Zone 1
Elastic Load
Balancing
Elastic Load
Balancing
Auto-Scaled
EC2
Instances
Auto-Scaled
EC2
Instances
AWS Region 1
AWS Region 2
AWS Region 3
AWS Region 4

10. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
#3: Automate Your Environment

11. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #2: Automate Your Environment
Removing manual processes to improve your
system's stability and consistency, and the
efficiency of your organization.
App
server
crashes
Replacement
automatically
launches

12. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
#4: Loosely Couple Your Components

13. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #4: Loosely Couple Your Components
Reduce interdependencies so that the
change or failure of one component
does not affect other components.
Load balancer reduces
dependency between
instances

14. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Cloud Architecture Principles
Loose coupling sets you free

15. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Cloud Architecture Principles
Loose coupling sets you free
• Design architectures with
independent components
• Design every component as a
black box
• Load balance clusters
• Separate application into
independent tiers

16. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Cloud Architecture Principles
Loose coupling sets you free
• Use queues to pass
messages between
components

17. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
#5: Design Services, Not Servers

18. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #5: Design Services, Not Servers
Managed services and serverless
architectures can provide greater reliability
and efficiency in your environment.
Amazon SQS
for message
queuing
AWS Lambda
for running simple
functions
Amazon SNS
for push notifications

19. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Simple Problem
Bucket with
photo uploaded
Bucket with
transformed
objects
• Monitor uploads
• Add watermark information
• Upload transformed photo into
bucket for wider team use

20. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Simple Problem
corporate data center
Proxy Servers
Monitor
Uploads
Upload
Photos
Process Servers
Enqueue
Processes
Solution:
Transform Objects
Development/Infrastructure Team:
• Infrastructure Planning
• Capacity
• Fault Tolerance
• Long-Term Utilization
• Burst Capacity
• 24/7 Monitoring:
• Capacity
• Health
• Security
• Etc.
• Over time:
• Upgrades and Migrations
• OS and Language Runtimes
Patched

21. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Why AWS Lambda?
AWS cloud
AWS
LambdaUpload
Photo
S3 bucket with
Raw Photo
Added Watermark
S3 bucket with
transformed Object
Put Event Transformed Photo

22. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
#6: Choose the Right Database Solutions

23. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #6: Choose the Right Database
Solutions
Match the technology to the workload:
Choose from an array of relational database
engines, NoSQL solutions, data warehousing
options, and search-optimized data stores.
Amazon Elasticsearch
for Search
Amazon Kinesis
for Streaming Data
Amazon DynamoDB
for NoSQL
Amazon RDS
for Relational
Database
Amazon Redshift
for Data Warehousing

24. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #6: Choose the Right Database
Solutions
• Don’t log clicks to RDBMS,
use NoSQL data store
• Don’t store images in
RDBMS, use object store
• Offload log files to scalable
object storage

25. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
#7: Removing Single Points of Failure

26. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Cloud Architecture Principles
“Everything fails, all the time.”
Werner Vogels, CTO, Amazon.com

27. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #7: Avoid Single Points of Failure
Implement redundancy where possible so
that single failures don't bring down an
entire system.
App servers
If one instance goes down,
another is available

28. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Cloud Architecture Principles
App
Server
Database
Server
Goal: Applications should
continue to function even if the
underlying application component
fails, communication is lost or
physical hardware fails, is
removed/replaced.

29. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Availability Zone 2
RDS DB
instancestandby
(multi-AZ)
RDS DB
instance
Availability Zone 1
Elastic Load
Balancing
Elastic Load
Balancing
RDS DB
instance
Auto-Scaled
EC2
Instances
Auto-Scaled
EC2
Instances

30. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Availability Zone 2
RDS DB
instancestandby
(multi-AZ)
RDS DB
instance
Availability Zone 1
Elastic Load
Balancing
Elastic Load
Balancing
Auto-Scaled
EC2
Instances
Auto-Scaled
EC2
Instances

31. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
#8: Optimize for Cost

32. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #8: Optimize for Cost
Ensure that your resources are sized appropriately,
that they scale in and out based on need, and that
you're taking advantage of different pricing options.

33. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Having decomposed into
small, loosely coupled,
stateless building blocks
You can now Scale out with ease
Having done that…

34. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Having decomposed into
small, loosely coupled,
stateless building blocks
We can also Scale back with ease
Having done that…

35. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Leveraging EC2’s Different Pricing Models
12am 6am 12pm 6pm 12am
Demand
Anticipated Weekday Demand
RI and On-Demand
Reserved Instances On-Demand Instances

36. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
12am 6am 12pm 6pm 12am
Demand
Anticipated Weekday Demand
RI, On-Demand, and Spot
Reserved Instances On-Demand Instances Spot Instances
Leveraging EC2’s Different Pricing Models

37. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon
Glacier
Lifecycle Policies
Amazon S3 lifecycle policies allow you to delete or move objects
based on age.
Amazon S3
Standard
Amazon S3
Standard -
Infrequent
Access
Delete
30 Days 60 Days 365 Days
Preview2.mp4 Preview2.mp4 Preview2.mp4

38. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
#9: Use Caching

39. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #9: Use Caching
Use caching to minimize
redundant data retrieval
operations.
Amazon S3
bucket with data
Caching with
CloudFront

40. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Availability Zone 2
RDS DB
instancestandby
(multi-AZ)
RDS DB
instance
Availability Zone 1
Elastic Load
Balancing
Elastic Load
Balancing
Auto-Scaled
EC2
Instances
Auto-Scaled
EC2
Instances

41. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Availability Zone 2
RDS DB
instancestandby
(multi-AZ)
RDS DB
instance
Availability Zone 1
Elastic Load
Balancing
Elastic Load
Balancing
Auto-Scaled
EC2
Instances
Auto-Scaled
EC2
Instances
Amazon
CloudFront
S3 Bucket

42. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
#10: Secure Your Infrastructure Everywhere

43. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Best Practice #10: Secure Your Infrastructure
Everywhere
AWS enables you to implement
security both at the perimeter and
within/between your resources.
Security group
Security group
Security group

44. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Cloud Architecture Principles
Build Security into every layer
• Encrypt data in transit and
rest between application tiers
• Enforce principle of least
privilege across every service
• Automatically rotate security
keys frequently

45. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Cloud Architecture Principles
Build Security
into every layer
Route
Table
Route
Table
Subnet 10.1.1.0/24 Subnet 10.1.10.0/24
Internet
Gateway
Virtual
Private
Gateway
Virtual
Router
VPC 10.1.0.0/16

46. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Availability Zone 2
RDS DB
instancestandby
(multi-AZ)
RDS DB
instance
Availability Zone 1
Elastic Load
Balancing
Elastic Load
Balancing
Auto-Scaled
EC2
Instances
Auto-Scaled
EC2
Instances
Amazon
CloudFront
AWS WAF

47. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Architecting for the Cloud: AWS Best Practices
1. Enable Scalability
2. Use Disposable Resources
3. Automate Your Environment
4. Loosely Couple Your Components
5. Design Services, Not Servers
6. Choose the Right Database Solutions
7. Avoid Single Points of Failure
8. Optimize for Cost
9. Use Caching
10. Secure Your Infrastructure Everywhere

48. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Thank You