This document summarizes the steps to scale an application hosted on Amazon Web Services (AWS) from a single instance to support millions of users. It begins with a single EC2 instance and expands to include load balancing, auto-scaling, database read replicas, content delivery networks, caching, and serverless architectures. The key aspects covered are separating concerns across multiple instances, adding redundancy, shifting load to managed services, and automating scaling.

1. ©2015, Amazon Web Services, Inc. or its affiliates. All rights reserved
Scale, baby, scale!
Julien Simon
Principal Technical Evangelist
Amazon Web Services
julsimon@amazon.fr
@julsimon

2. So let’s start from
day 1, user 1 (you)

3. Day 1, user 1
• A single Amazon EC2
instance, with full stack on
this host
– Web app
– Database
– Management
– And so on…
• A single Elastic IP address
• Amazon Route 53 for DNS
Amazon
EC2
instance
Elastic IP
address
Amazon
Route 53
User

4. “We’re gonna need a bigger box”
• “Scale up”: simplest approach
• Can now leverage PIOPS
• High I/O instances
• High memory instances
• High CPU instances
• High storage instances
• Easy to change instance sizes
• Will hit a wall eventually
c4.8xlarge
m4.2xlarge
t2.micro

5. Day 1, user 1
• We could potentially get
to a few hundred to a few
thousand depending on
application complexity
and traffic
• No failover
• No redundancy
• Too many eggs in one
basket
EC2
instance
Elastic IP
address
Amazon
Route 53
User

6. AWSome Days 3.8

7. 100 users
First, let’s separate out
our single host into
more than one:
• Web
• Database
Use Amazon RDS to make
your life easier
Web
instance
Elastic IP
address
RDS DB
instance
Amazon
Route 53
User

8. 1000 users
Next, let’s address our
lack of failover and
redundancy issues:
• Elastic Load
Balancing (ELB)
• Another web instance
In another Availability
Zone
• RDS Multi-AZ
Web
Instance
RDS DB Instance
Active (Multi-AZ)
Availability Zone Availability Zone
Web
Instance
RDS DB Instance
Standby (Multi-AZ)
ELB
Balancer
Amazon
Route 53
User

9. Scaling this horizontally and
vertically will get us pretty far
(tens to hundreds of
thousands)

10. 10k-100k users
RDS DB Instance
Active (Multi-AZ)
Availability Zone Availability Zone
RDS DB Instance
Standby (Multi-AZ)
ELB
Balancer
RDS DB Instance
Read Replica
RDS DB Instance
Read Replica
RDS DB Instance
Read Replica
RDS DB Instance
Read Replica
Web
Instance
Web
Instance
Web
Instance
Web
Instance
Web
Instance
Web
Instance
Web
Instance
Web
Instance
Amazon
Route 53
User

11. This will take us pretty far, but
we care about performance
and efficiency, so let’s
improve further

12. RDS DB Instance
Active (Multi-AZ)
Availability Zone
ELB
Balancer
Amazon
Route 53
User
Let’s lighten the load on our
web and database instances
Shift some load around
Web Instances

13. RDS DB Instance
Active (Multi-AZ)
Availability Zone
ELB
Balancer
Amazon S3
Amazon
CloudFront
Amazon
Route 53
User
Let’s lighten the load on our
web and database instances:
• Move static content from
the web instance to
Amazon S3 and Amazon
CloudFront
• We could even host the
site directly in S3 (aka
“S3 static website”)
Shift some load around
Web Instances

14. Shift some load around
Let’s lighten the load on our
web and database instances:
• Move static content from the
web instance to Amazon S3
and Amazon CloudFront
• Move session/state and
DB caching to Amazon
ElastiCache or Amazon
DynamoDB
RDS DB Instance
Active (Multi-AZ)
Availability Zone
ELB
Balancer
Amazon S3
Amazon
CloudFront
Amazon
Route 53
User
ElastiCache
DynamoDB
Web Instances

15. Shift some load around
Let’s lighten the load on our
web and database instances:
• Move static content from the
web instance to Amazon S3
and Amazon CloudFront
• Move session/state and DB
caching to ElastiCache or
DynamoDB
• Move dynamic content
from the ELB balancer to
Amazon CloudFront
RDS DB Instance
Active (Multi-AZ)
Availability Zone
ELB
Balancer
Amazon S3
Amazon
CloudFront
Amazon
Route 53
User
ElastiCache
DynamoDB
Web Instances

16. Now that our web tier is
much more lightweight…

17. Auto Scaling!
Resize server farms automatically
based on monitoring metrics
Spot instances!
Bid on unused EC2 capacity
(typically at 80% discount)

18. 500k users
Availability Zone
Amazon
Route 53
User
Amazon S3
Amazon
CloudFront
Availability Zone
ELB
Balancer
DynamoDB
RDS DB Instance
Read Replica
Web
Instance
Web
Instance
Web
Instance
ElastiCache RDS DB Instance
Read Replica
Web
Instance
Web
Instance
Web
Instance
ElastiCacheRDS DB Instance
Standby (Multi-AZ)
RDS DB Instance
Active (Multi-AZ)

19. AWSome Days 3.8

20. There are more
improvements to be made
and we could get much
higher, but do we really
want to manage all these
instances ?

21. Werner Vogels, CTO, Amazon.com
AWS re:Invent 2015

22. Use the Force, Luke!
Managed services
+ AWS Lambda
= Serverless architecture

23. AWSome Days 3.8

24. Upcoming book on AWS Lambda
Written by AWS Technical Evangelist
Danilo Poccia
Early release available at:
https://www.manning.com/books/aws-
lambda-in-action

25. ©2015, Amazon Web Services, Inc. or its affiliates. All rights reserved
Thank You !
Julien Simon
julsimon@amazon.fr
@julsimon