This document provides an overview and summary of AWS storage services that can be used for migrating data to AWS. It discusses AWS Snowball and Snowmobile appliances that can physically move large amounts of data to AWS storage services like S3. It also describes the AWS Storage Gateway, which allows on-premises applications to access AWS storage using standard storage protocols. Additional services covered include Amazon Kinesis Firehose for loading streaming data, AWS Direct Connect for private connectivity, and AWS Migration Hub and Application Discovery Service for discovery and tracking of servers and databases during migration.
3. Data
Data in the cloud has gravity
Deliver new
insights
(data lakes, analytics)
Accelerate
innovation
(active archive, IoT,
Artificial Intelligence)
Realize benefits
(cost, management, scale)
Build or migrate
an application
5. Compliance
Industry
certifications
Lockable with audit
trails
Secure
Enterprise
Applications
Easier lift-and-shift
migrations
Integrated with
major vendors
Fully managed
infrastructure
Active
Archive
Media workflows
Tape replacement
Public Sector,
FinServ,
Healthcare/Life
Sciences
Databases &
Analytics
Tailored database
or Hadoop
workloads
Bespoke database
lift-and-shift
projects
Backup and
Restore
Non-disruptive
Easy place to start
Integrated with all
major vendors
Data Lakes
& IoT
400% faster
queries
Built for
streaming data
Optional data
visualization
Common storage workloads on AWS
6. Why AWS Storage?
The best reliability and largest scale The most complete portfolio
The most data movement choices The most comprehensive
support and consulting
More than twice the partners
The most secure,
compliant, and auditable
7. Complete set of building blocks
Data movement Data security
and management
File
Storage
Block
Storage
Object
Storage
Archival
Storage
Encryption
Access Controls
Monitoring and Metrics
Audit Trails
Automation
Serverless Computing
Data Discovery and
Protection
Data Visualization
Physical Appliances
Hybrid Storage
Private Networks
File Data
WAN Acceleration
Third-party
Applications
Streaming Data
8. AWS storage services
Data movement
OnlineOffline
Data security
and management
Amazon
EFS
Amazon
EBS
Amazon
S3
Amazon
Glacier
AWS KMS
AWS IAM
Amazon CloudWatch
AWS CloudTrail
AWS CloudFormation
AWS Lambda
Amazon Macie
Amazon QuickSight
AWS Snow Family
AWS Storage Gateways
AWS Direct Connect
Amazon EFS File Sync
Amazon S3 Transfer
Acceleration
Third-party
Applications
Amazon Kinesis Firehose
12. 2017 Gartner Magic Quadrant
- Gartner Magic Quadrant for Public Cloud Storage Services, Worldwide
Raj Bala, Arun Chandrasekaran, John McArthur, July 24, 2017
“AWS sets the boundaries in the market
for public cloud storage services
by which all other vendors operate.”
13. Reliability and Scale
“…the scale at which AWS operates its public
cloud storage services dwarfs the other vendors
in this Magic Quadrant.”
- Gartner Magic Quadrant for Public Cloud Storage Services, Worldwide
Raj Bala, Arun Chandrasekaran, John McArthur, July 24, 2017
For example: Amazon S3 holds trillions of objects and
regularly peaks at millions of requests per second
TIME
OBJECTS
14. Amazon S3
Analyze
Store
Collect
Built for:
backup and restore, data lakes, analytics, cloud-native applications
• More than a decade of experience and continuous innovation
• Only AWS has the infrastructure to place storage near workloads
• Only AWS gives storage admins granular object controls
• Only AWS moves data in so many varied ways
• Only AWS storage can help address CISO concerns
• Only AWS can analyze and recommend cost savings
• Only AWS can accelerate application performance up to 400%
• Only AWS offers inventory and visualization across entire datasets
• Only AWS supports queries across structured and unstructured data
15. Object storage classes
S3 Standard GlacierS3 Standard -
Infrequent Access
Active data
Milliseconds
$0.023/GB/mo
Archive data
Minutes to Hours
$0.004/GB/mo
Infrequently accessed data
Milliseconds
$0.0125/GB/mo
Automated Lifecycle Policies
16. Amazon Glacier
Cost-
effective
Secure
Durable
Built for:
Active archive, tape replacement, regulatory compliance
• Certifications supporting nearly any regulatory compliance program
• Locking, encryption, audit and alerting tools to prevent tampering
• Built on the most reliable global infrastructure
• Withstands multiple facility failures
• Replication options across global regions
• Designed for archives and backup
• Expedited retrievals in minutes, bulk retrievals in hours
• Opens archives to analytics applications
17. Amazon S3 and Glacier Durability
Designed for
99.999999999% durability
GlacierS3 Standard S3 - IA
OR:
99.999% durability
99.99% durability
Traditional model with two
copies on one site
Traditional model with copies on
two sites
S3 – IA OZ
18. Amazon Object Storage Availability and Durability
“Zones”
Or worse, this:
AWS Region
This:
Availability Zone
Availability Zone
Availability Zone
Not this:
“Region”
20. Security-as-a-Service for 4000 customers
using 25 PB and growing 110% per year
Colocate simply not agile enough or cost effective
• Built an Amazon S3 data lake and avoided
$1.6M CAPEX - in the first year alone
• Stress-tested 100x larger load with zero CAPEX
• 4x better “I/O per $” ratio
• Gained new insights into their customers
through Amazon S3 data management
capabilities
• No 40-Gbps network infrastructure worries
“AWS storage is
Fully redundant, multi-region,
Mmore secure, and faster
at less than half the cost.”
- Paul Fisher
Technical fellow
S 3 D A T A L A K E
Data Lakes
21. Threat analysis company ingesting
and analyzing 50 TB daily
Right-sizing clusters cost weeks and lost data
• Saved 95% through re-architecting to a “hot”
index on Amazon EBS with an analytics data
lake on Amazon S3
• Amazon EBS shortened indexing times from
weeks to hours while cutting OPEX
• Now getting consistent 1–3 sec. search
response times across 5 PB of growing data in
Amazon S3
• Managing 1 billion Amazon S3 objects and
2,500 instances with just six engineers
“AWS storage completely changed our
business operations, time to market and
manpower. EBS volumes cut our cluster
indexing times from weeks to hours. Moving
data into Amazon S3 saved us 95% and our
data lake now outperforms our clusters—the
harder we push it the faster it gets for
extremely large datasets. We simply could not
do this anywhere else.”
- Gene Stevens
CTO and cofounder
Data Lakes
A m a z o n S 3 D A T A L A K E
25. Databases and analytics
Global broadband service operator processing
17 TB of daily device data streams at 200 MB/s
Modifying Kafka clusters required an
8 hours resync every time
• Moved from instance stores to EBS volumes
• Cut storage costs by 25%
• Cut production cluster node count by 33%
• Dropped resync times to 20 minutes
“Our AWS service use is about making the
necessary easy. Storage should be as boring
as possible—it should just work. Amazon
EBS makes it trivial to do things that were
impractical before, driving experimentation,
creativity, and faster delivery.”
- Daniel Woodlins
Software engineer
27. Amazon EFS
Scalable
Simple
Elastic
Built for:
Web serving, content management, media and entertainment workflows, home
directories, container storage, big data, and analytics
• Share files between EC2 instances in minutes
• True file system interface with file system semantics
• Fully managed – no capacity planning surprises
• Pay-as-you-go consumption and pricing
• Automatically grows and shrinks
• Much lower TCO than DIY or third-party workarounds
• Consistent performance even as data grows
29. Newly acquired streaming media product
depended on a local file server
Had to launch at global scale in 90 days – with
minimal changes
• DIY was too complex and took too long
• Lift-and-shift to Amazon EFS took 2 hours
• EFS with EC2 autoscaling met global scale
agility needs
• Seamless integration between partner
application and existing AWS systems
• Post-mortem TCO analysis showed that EFS
was still the best choice
Enterprise applications
“Good, fast, and cheap. We picked two and
got all three with Amazon EFS. It gave us
the agility to deliver a new product on
schedule, eliminated scale and
performance concerns, and operates below
our
OPEX expectations.”
- Chris DeAcosta
Sr. director software engineering
30. Enterprise applications
Builds 3d digital maps relying on 28 TB of
waypoints generated daily
Unreliable on-premises repository and
high maintenance DIY cloud version
• Amazon EFS dropped infrastructure provisioning
time from 90 days to 7
• Now handling 800,000 daily file transfers up to
38% faster with zero failures
• Seamless JFrog workflow integration
• Gained high availability at no extra cost
• Also tiering JFrog backups into Amazon S3 and
Amazon Glacier
Prior to Amazon EFS, we experienced
timeouts for up to 10% of uploads over
100 MB. Now, all of the JFrog build
artifacts (from infrastructure-as-code
components to Docker images) are in one
place, and we’ve increased large file
transfer speeds by 38%.”
- Suresh Prem, Murty Chitti,
and Rajesh Sivaraman
System engineers