Scaling Up To and Beyond 10M Users

© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Confidential and Trademark© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Confidential and Trademark
Giorgio Bonfiglio (bonfigg@amazon.com)
Sr. TAM - Enterprise Support - AWS
Scaling up to 10M users… and beyond!
(how to be ready in case your application gets as
successful as you were hoping)
Thursday, Sept 20th 2018

Now that’s a lot
of things to read!
This is NOT
where we
want to start!

It’s not the single thing that
fixes everything.

What do we need first?

Some basics…

AWS Global Infrastructure
Region (# of Zones)
Coming Soon!

AWS Global Infrastructure
Region (# of Zones)
Coming Soon!
125+ Edge
Locations

ENTERPRISE
APPS
DEVELOPMENT & OPERATIONSMOBILE SERVICESAPP SERVICESANALYTICS
Data
Warehousin
g
Hadoop
/Spark
Streaming
Data Collection
Machine
Learning
Elastic
Search
Virtual
Desktops
Sharing &
Collaboratio
n
Corporate
Email
Backup
Queuing &
Notification
s
Workflo
w
Search
Email
Transcoding
One-click
App
Deployment
Identity
Sync
Single
Integrated
Console
Push
Notification
s
DevOps Resource
Management
Application
Lifecycle
Management
Containers
Triggers
Resource
Templates
TECHNICAL &
BUSINESS
SUPPORT
Account
Managemen
t
Support
Professional
Services
Training &
Certificatio
n
Security
& Pricing
Reports
Partner
Ecosystem
Solutions
Architects
MARKETPLACE
Business
Apps
Business
Intelligenc
e
Databases
DevOps
Tools
Networkin
g
Securit
y
Storage
Regions
Availability
Zones
Points of
Presence
INFRASTRUCTURE
CORE SERVICES
Compute
VMs, Auto-scaling,
& Load Balancing
Storage
Object, Blocks,
Archival,
Import/Export
Databases
Relational, NoSQL,
Caching, Migration
Networkin
g
VPC, DX, DNS
CDN
Access
Control
Identity
Management
Key
Management
& Storage
Monitoring
& Logs
Assessment
and reporting
Resource &
Usage Auditing
SECURITY & COMPLIANCE
Configuration
Compliance
Web
application
firewall
HYBRID
ARCHITECTURE
Data Backups
Integrated
App
Deployments
Direct
Connect
Identity
Federation
Integrated
Resource
Management
Integrated
Networkin
g
API
Gateway
IoT
Rules
Engine
Device
Shadows
Device
SDKs
Registry
Device
Gateway
Streaming
Data Analysis
Business
Intelligence
Mobile
Analytics
Most Robust, Fully-Featured Technology Infrastructure Platform

AWS building blocks
Inherently highly scalable, available and
fault-tolerant services
Highly scalable,
available with the right
architecture
a Amazon CloudFront
a Amazon Route 53
a Amazon S3
a Amazon DynamoDB
a Elastic Load Balancing
a Amazon EFS
a AWS Lambda
a Amazon SQS
a Amazon SNS
a Amazon SES
a AWS Step Functions
a …
4 Amazon EC2
4 Amazon EBS
4 Amazon RDS
4 Amazon VPC

Considerations
Rex Boggs
https://creativecommons.org/licenses/by-nd/2.0/

“Many decisions are
reversible, two-way
doors.”
-- Jeff Bezos

Build Measure
Learn

So let’s start from…

You

1 User
Amazon
EC2
instance
Elastic IP
User
Amazon
Route 53

Lightsail: the easiest way to get started on AWS
• Choose from five plans that include bundled
compute, storage, and networking
• Benefit from a low, predictable price
• Spin up a fully configured server in seconds
• Manage from the intuitive Lightsail console
• Scale with access to AWS services
• Automate with Lightsail API & CLI

“We’re gonna need a bigger box”
• Simplest approach
• High I/O instances
• High memory instances
• High CPU instances
• High storage instances
• Easy to change instance sizes
• Will hit an endpoint eventually
c4.8xlarge
m4.2xlarge
t2.micro

1 User
• No failover
• No redundancy
• Too many eggs in one
basket
EC2
Instance
Elastic IP
User
Amazon
Route 53

Users >1

Users > 1
Web
Instance
Database
Instance
Elastic IP
User
Amazon
Route 53

Self-managed Fully managed
Amazon EC2 Amazon
DynamoDB
Amazon RDS Amazon
Redshift
Database options

• MySQL or Postgres compatible
• Automatic storage scaling (up to 64 TB)
• Up to 15 read-replicas
• Continuous backups to Amazon S3
• 6-way replication across 3 zones
• Comes into a serverless flavor!
• (available soon: Multi-Master)
Amazon Aurora

To NoSQL, or not to NoSQL?

Start with SQL databases

Why start with SQL?
• Established and well-worn technology
• Lots of existing code, communities, books, and tools
• You aren’t going to break SQL DBs in your first millions
of users
• No, really, you won’t*
• Clear scalability patterns
*Unless you are doing something SUPER peculiar with the data or you have MASSIVE amounts of it...
…but even then SQL will have a place in your stack.

AH HA! You said
“massive
amounts”

Why might you need NoSQL?
• > 5TB data in first year
• Super low-latency applications
• Metadata-driven datasets
• Highly non-relational data
• Need schema-less data constructs*
• Rapid ingest of data (thousands of records/sec)
• Massive amounts of data (in the TB range)
*Need != “It’s easier to do dev without schemas”

Users >100

Users >100
Web
instance
Elastic IP
Amazon
RDS DB
instance
User
Amazon
Route 53

Users >1000

Users >1000
Web
Instance
Amazon RDS DB Instance
Active (Multi-AZ)
Availability Zone Availability Zone
Web
Instance
Standby (Multi-AZ)
Load
balancer
User
Amazon
Route 53

Load Balancers
Classic Load BalancerApplication Load Balancer Network Load Balancer

Application Load Balancer
• Highly available
• Health checks
• Session stickiness
• Monitoring / logging
• Content-based routing
• Container-based apps
• WebSockets
• HTTP/2
Recommended

Network Load Balancer
• Layer 4
• Elastic / Static IP
• Preserve source
• Extreme performance
• Ultra-low latency
• Volatile workloads

horizontally
vertically

Users >100,000

Users >100,000s
Amazon RDS DB
Instance
Active (Multi-AZ)
Availability Zone Availability Zone
Amazon RDS DB
Instance Standby
(Multi-AZ)
Amazon RDS DB
Instance Read
Replica
Amazon RDS DB
Instance Read
Replica
Amazon RDS DB
Instance Read
Replica
Amazon 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

Active (Multi-AZ)
Availability Zone
Load
balancer
Amazon S3
Amazon
CloudFront
Amazon
Route 53
User
Shift some load around
Web Instances
Availability Zone
Web Instances
Standby (Multi-AZ)

• Object-based storage
• Highly durable
• Great for static assets
• “Infinitely scalable”
• Objects up to 5 TB in size
• Encryption at rest and in
transit
Amazon Simple Storage Service (S3)

• Cache content for faster
delivery
• Lower load on origin
• Dynamic and static content
• Streaming video
• Custom SSL certificates
• Low TTLs (as short as 0
seconds)
Amazon CloudFront

Amazon CloudFront
Response
Time
ServerLoad
Response
Time
Serve
r
Load
Response
Time
Server
Load
No CDN CDN for Static
Content
CDN for Static
& Dynamic
Content
0
10
20
30
40
50
60
70
80
8:00
A
M9:00
A
M
10:00
AM
11:00
AM
12:00
PM1:00
P
M2:00
P
M3:00
P
M4:00
P
M5:00
P
M6:00
P
M7:00
P
M8:00
P
M9:00
P
M
VolumeofData
Delivered(Gbps)

Active (Multi-AZ)
Availability Zone
Load
balancer
Amazon S3
Amazon
CloudFront
Amazon
Route 53
User
Web Instances
Amazon
ElastiCache

Amazon ElastiCache
• Managed Memcached or Redis
• Scale from one to many nodes
• Self-healing
• Single digit ms speeds
• Multi-AZ possible with Redis

Active (Multi-AZ)
Availability Zone
Load
balancer
Amazon S3
Amazon
CloudFront
Amazon
Route 53
User
Amazon
ElastiCache
Amazon
DynamoDB
Web Instances

Amazon DynamoDB
• Managed NoSQL database
• Provisioned throughput
• Fast, predictable performance
• Fully distributed, fault tolerant
• JSON support
• Items up to 400 KB
• Time-to-live (TTL)
• Streams and Triggers
• Can scale automatically!

Amazon DynamoDB Accelerator
DynamoDB
Your Applications
DynamoDB Accelerator
• Read performance and scale: Microseconds
response times at millions of reads/sec from
single DAX cluster
• Lower costs: Reduce provisioned read capacity
for DynamoDB tables for tables with hot data

Now that our web tier is
much more lightweight,
we can revisit the beginning
of our talk…

Auto Scaling!

Sunday Monday Tuesday Wednesday Thursday Friday Saturday
Typical weekly traffic to Amazon.com

Sunday Monday Tuesday Wednesday Thursday Friday Saturday
Typical weekly traffic to Amazon.com
Provisioned capacity

November
November traffic to Amazon.com

November

76%
24%
November

November

Auto Scaling
lets you do this!

Automatic resizing of instance fleets
Define min/max pool sizes
CloudWatch metrics drive scaling
On-Demand and/or Spot Instances
aws autoscaling create-auto-scaling-group
--auto-scaling-group-name MyGroup
--launch-configuration-name MyConfig
--min-size 4
--max-size 200
--availability-zones eu-west-2a, eu-west-2b
Auto Scaling

Users > 500,000+
Availability Zone
Amazon
Route 53
User
Amazon S3
Amazon
CloudFront
Availability Zone
Load
balancer
Amazon
DynamoDBAmazon RDS DB
Instance Read
Replica
Web
Instance
Web
Instance
Web
Instance
Amazon
ElastiCache
Amazon RDS DB
Instance Read
Replica
Web
Instance
Web
Instance
Web
Instance
Amazon
ElastiCache
Amazon RDS DB
Instance Standby
(Multi-AZ)
Amazon RDS DB
Instance
Active (Multi-AZ)

Use automation

AWS application management solutions
Convenience Control
Higher-level services Do it yourself
AWS
Elastic Beanstalk
AWS
OpsWorks
AWS
CloudFormation
Amazon EC2

Users >500,000+
• Monitoring, metrics, and logging
• What are customers saying?
• Try to squeeze as much performance
out of each service/component

AGGREGATE
LEVEL
METRICS
LOG
ANALYSIS
EXTERNAL
SITE PERFORMANCE
HOST
LEVEL
METRICS

CloudWatch percentiles

There are further
improvements to be made in
breaking apart our
web/app layer

The Monolithic Architecture
User Interface
Business Logic
Data Access

SOA
What does this mean?

The Service-Oriented Architecture
Presentation Tier Logic Tier
Data Tier

SOAing
Move services into their own tiers
• Treat them separately
• Scale them independently
It offers flexibility and greater
understanding of each component

Loose coupling + Serverless = winning
DON’T REINVENT THE WHEEL
• Email
• Queuing
• Transcoding
• Search
• Databases
• Monitoring
• Metrics
• Logging
• Compute
Amazon SQSAmazon SNS
Amazon Elastic
Transcoder
AWS Step
Functions
Amazon SESAWS Lambda
Amazon Elastic
Search
Amazon API
Gateway

• Reliable (Multi-AZ)
• Scalable (unlimited messages)
• Secure (queue authentication)
• Simple (simple APIs)
Application Services – Amazon SQS
Amazon
SQS
messages
Get
message
Instance
Put
message
Instance
Amazon SNS Topic
Publish
notification
Queue is subscribed
to topic

Event Driven Compute – AWS Lambda
• Functions triggered by events
• Node.js, Java, Python, and C#
• Managed
• Implicit scaling
S3 bucket
Lambda
Push: event
notification
DynamoDB
Pull: DynamoDB
Stream
Amazon
Kinesis
Pull:
Amazon Kinesis
stream

Loose coupling sets you free!
The looser they're coupled, the bigger they scale
• Independent components
• Design everything as a black box
• Decouple interactions
• Favor services with built-in redundancy and scalability
• Don’t build your own!
S3 bucket
Lambda
Push: event
notification
DynamoDB
Pull: DynamoDB
Stream
Amazon
Kinesis
Amazon
SQS
messages
Get
message
Instance
Put
message
Instance
Amazon SNS Topic
Publish
notification
Queue is subscribed
to topic
Pull:
Amazon Kinesis
stream

Serverless Web application
Data stored in
Amazon
DynamoDB
Dynamic content
in AWS Lambda
Amazon API
Gateway
Browser
Amazon
CloudFront
Amazon
S3

The Microservices Architecture

Microservices in AWS

AWS X-Ray
• Identify performance bottlenecks and errors
• Pinpoint issues to specific service(s) in your
application
• Identify impact of issues on users of the application
• Visualize the service call graph of your application

Visualize service call graph

Users >1,000,000

Users >1 million+
Reaching a million and above is going to require some bit
of all the previous things:
• Multi-AZ
• Elastic Load Balancing between tiers
• Auto Scaling
• Service oriented architecture (SOA)
• Serving content smartly (Amazon S3/CloudFront)
• Caching off DB
• Moving state off tiers that auto scale

Users >1 million+
Amazon RDS DB
Instance
Active (Multi-AZ)
Availability Zone
Amazon RDS DB
Instance Read
Replica
Amazon RDS DB
Instance Read
Replica
Web
Instance
Web
Instance
Web
Instance
Web
Instance
User
Amazon S3
Amazon
DynamoDB
Amazon SQS
ElastiCache
Worker
Instance
Worker
Instance
Amazon
CloudWatch
Internal App
Instance
Internal App
Instance Amazon SES
AWS Lambda
Amazon
Route 53
Amazon
CloudFront
Load
balancer

Users >1 million+
Amazon RDS DB
Instance
Active (Multi-AZ)
Availability Zone
Amazon RDS DB
Instance Read
Replica
Amazon RDS DB
Instance Read
Replica
Web
Instance
Web
Instance
Web
Instance
Web
Instance
User
Amazon S3
Amazon
DynamoDB
Amazon SQS
ElastiCache
Worker
Instance
Worker
Instance
Amazon
CloudWatch
Internal App
Instance
Internal App
Instance Amazon SES
AWS Lambda
Amazon
Route 53
Amazon
CloudFront
Load
balancer
Amazon
DynamoDB
Amazon API
Gateway

The next big steps

Users >5 million - 10 million
Database Issues?
How can you solve it?
• Federation: splitting into multiple DBs based on function
• Sharding: splitting one dataset up across multiple hosts
• Moving some functionality to other types of DBs (NoSQL, Graph)

Database federation
• Split up databases by function/purpose
• Harder to do cross-function queries
• Essentially delays sharding/NoSQL
• Won’t help with single huge functions/tables
Forums DB
Users DB
Products
DB

Sharded horizontal scaling
• More complex at the application layer
• No practical limit on scalability
• Operation complexity/sophistication
• Shard by function or key space
• RDBMS or NoSQL
User ShardID
002345 A
002346 B
002347 C
002348 B
002349 A
CBA

Shifting functionality to NoSQL
• Similar in a sense to federation
• NoSQL vs. SQL
• Leverage managed services like
DynamoDB
Some use cases:
• Leaderboards/scoring
• Rapid ingest of clickstream/log data
• Temporary data needs (cart data)
• “Hot” tables
• Metadata/lookup tables
DynamoDB

A quick review

A quick review
• Multi-AZ your infrastructure
• Make use of self-scaling services: ALB, Amazon S3, AWS
Lambda, Amazon SNS, Amazon SQS, AWS Step Functions,
etc
• Build in redundancy at every level
• Start with SQL. Seriously.
• Cache data both inside and outside your infrastructure
• Use automation tools in your infrastructure

A quick review continued
• Make sure you have good metrics/monitoring/logging
• Split tiers into individual services (SOA)
• Use Auto Scaling once you’re ready for it
• Don’t reinvent the wheel
• Move to NoSQL if and when it makes sense

10+ million users!

Time to Go Global!

User >10 million
• More fine-tuning of your application
• More SOA of features/functionality
• Going from Multi-AZ to multi-region
• Deep analysis of your entire stack

Amazon CloudFront Global Content Delivery Network
136 PoPs (125 Edge Locations + 11 Regional Edge Caches)

Amazon CloudFrontAWS Lambda
Lambda@Edge
Lambda@Edge

Lambda@Edge
Content-based routing
to any origin
Response generation
• Generate custom responses at the
edge
• Read and write access to headers,
query string, and cookies
Network calls
• Origin events
• Viewer events

Transparent global expansion
Region A
customers
Region A
deployment
Region B
customers Region B
deployment
https://saas.example.com

Lambda@Edge
Origin
AWS Location
AWS Location
AWS Location
AWS Location
AWS Location
AWS Location
Compute
Storage
Database

DynamoDB Global Tables
Fully managed, multimaster, multiregion
database
Build high-performance, globally distributed
applications
Low latency reads & writes to locally available
tables
Disaster proof with multi-region redundancy
Easy to set up, and no application rewrites
required

To infinity...

Next steps?
READ!
aws.amazon.com/documentation
aws.amazon.com/architecture
aws.amazon.com/solutions
aws.amazon.com/quickstart
START USING AWS:
aws.amazon.com/free

You’re not alone!
forums.aws.amazon.com
aws.amazon.com/premiumsupport
Your (Technical) Account Manager
A Solutions Architect

Thank you!
Giorgio Bonfiglio – bonfigg@amazon.com

Scaling Up To and Beyond 10M Users

Recomendados

Recomendados

Mais conteúdo relacionado

Mais procurados

Mais procurados (20)

Semelhante a Scaling Up To and Beyond 10M Users

Semelhante a Scaling Up To and Beyond 10M Users (20)

Mais de Amazon Web Services

Mais de Amazon Web Services (20)

Scaling Up To and Beyond 10M Users