In this session, we explore the world’s first cloud-scale file system and its targeted use cases. Learn about Amazon EFS features and benefits, how to identify applications that are appropriate for use with Amazon EFS, and details about its performance and security models. The target audience is security administrators, application developers, and applications owners who operate or build file-based applications.

1. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
AWS re:INVENT
Deep Dive on Amazon Elastic File
System (Amazon EFS)
Edward Naim, Head of Product
Darryl Osborne, Solutions Architect
N o v e m b e r 2 8 , 2 0 1 7
S T G 3 0 7

2. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Your journey to Amazon EFS in four
phases
Phase 1:
Choose the right
storage solution
Phase 2:
Test and
optimize
Phase 3:
Ingest
Phase 4:
Run

3. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Phase 1:
Choose the right storage solution

4. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
What do you think about
when choosing a storage solution?
EconomicsStorage type Features
and performance

5. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
File
Three types of storage
ObjectBlock

6. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
File
Three types of storage
Data stored as files in a
directory hierarchy
Shared over a network

7. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Block
Three types of storage
Data stored as blocks on a
disk or disks
Locally attached

8. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Object
Three types of storage
Data is stored as an object
that’s identified by a key in a
flat space
Simple API to get and put data
based on key

9. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Why is file storage so popular?
Works natively with operating systems
Provides shared access while providing consistency guarantees and
locking functionality
Provides hierarchical namespace

10. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
File
How does performance compare?
Object
Block
Latency
Throughput

11. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Before Amazon EFS… DIY file storageAZ-a
Clients
Storage
volumes
AZ-b
Storage
volumes
File server
File server

12. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon EFS
A fully managed file service

13. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Key features of Amazon EFS
ScalableSimple Elastic
Highly available and durable

14. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Key features of Amazon EFS
ScalableSimple Elastic
Highly available and durable

15. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Key features of Amazon EFS
ScalableSimple Elastic
Highly available and durable

16. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Key features of Amazon EFS
ScalableSimple Elastic
Highly available and durable

17. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Key features of Amazon EFS
ScalableSimple Elastic
Highly available and durable

18. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Designed for a wide spectrum of needs
Scale-out jobs Metadata-intensive jobs
Analytics
Media
workflows
Web serving
Content management
Database backups
Enterprise apps
and messaging
Dev tooling
Home directories
Container storage
Low latency and serial I/OHigh throughput and parallel I/O

19. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon EFS customers
Web serving + content management
Media and entertainment workflowsDatabase backups
Big data and analytics
Enterprise applications
Developer tools Home directories Container storage

20. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Security model and security features
Control
network traffic
using Amazon VPC security
groups and network ACLs
Control file and
directory access
using POSIX
permissions
Control administrative
access (API access)
using AWS IAM
(action-level and
resource-level
permissions)
Encrypt
data at rest
using keys
managed in
AWS KMS

21. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Access EFS from various environments
Amazon
EFS
Amazon
EC2
Corporate data
center
VMware Cloud
on AWS

22. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon EFS economics
$0.30/GB-Month (US Regions)
$0.33/GB-Month (EU Ireland)
$0.36/GB-Month (EU Frankfurt)
$0.36/GB-Month (AP Sydney)
No minimum commitments
or up-front fees
No need to provision
storage in advance
No other fees, charges,
or billing dimensions

23. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Before EFS…costs of DIY file storageAZ-a
Clients
Storage
volumes
AZ-b
Storage
volumes
File server
File server
Amazon EC2
instance costs
Inter-AZ data
transfer costs
Amazon EBS
volume costs

24. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
TCO example: For storing 500 GB,
Amazon EFS is 70% less than DIY
Amazon EFS cost: (500GB * $0.30/GB-month) = $150 per month
For DIY, you might provision 600 GB of Amazon EBS (i.e., ~85% utilization):
Storage (2x 600 GB EBS gp2 volumes): $120 per month
Compute (2x m4.xlarge instances): $290 per month
Inter-AZ data transfer costs (est.): $130 per month
Total $540 per month

25. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Where is Amazon EFS available today?
• US West (Oregon)
• US East (N. Virginia)
• US East (Ohio)
• EU (Ireland)
• EU (Frankfurt)
• Asia Pacific (Sydney)
More coming soon!

26. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Data movement
OnlineOffline
Data security
and management
Complete set of data building blocks
Amazon
EFS
Amazon
EBS
AWS Snow Family
AWS Storage Gateway
Family
AWS Direct Connect
Amazon EFS File Sync
Amazon S3
Transfer Acceleration
Storage Partners
Amazon Kinesis
Data Streams
Amazon Kinesis
Video Streams
Amazon
S3
Amazon
Glacier
AWS KMS
AWS IAM
AWS CloudWatch
AWS CloudTrail
AWS CloudFormation
AWS Lambda
Amazon Macie
AWS QuickSight

27. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Introducing EFS File Sync
Fast
Up to 5x faster than
standard Linux copy tools
Sync data from existing file systems into Amazon EFS file systems
Secure
Encrypted parallel data
transfer to AWS
Simple
Set up and manage easily
from the AWS Console
file systems from on-premises to EFS
DIY in-cloud file systems to EFS
EFS file systems between AWS Regions
Use EFS File Sync to copy…

28. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Phase 2:
Test and optimize

29. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
What do you think about?
Performance testing
and optimization
Solution
architecture
Functional testing

30. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon EFS architecture
NFS clients Mount target
Mount target
Mount target
AZ-aAZ-bAZ-c
DNSName
Amazon EFS
File System
NFS clients
NFS clients

31. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
How to manage file systems
Create a file system
Create and manage mount targets
Tag a file system
Delete a file system
View details on a file system
API

32. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Functional testing
POSIX compliant & compliant with
NFS 4.0 & 4.1
But you always need to test your
application

33. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Choose the performance mode
best suited to your workload
Mode What it is for Advantages Trade-offs When to use
General
purpose
(default)
Latency-
sensitive
applications and
general-purpose
workloads
Lowest latencies
for file operations
Limit of 7k
ops/sec
Best choice for
most workloads
Max I/O
Large-scale and
data-heavy
applications
Virtually
unlimited ability
to scale out
throughput/IOPS
Slightly
higher
latencies
Consider for
large scale-out
workloads

34. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Monitor PercentIOLimit
CloudWatch metric
PercentIOLimit shows how close a file system is to reaching the I/O
limit of the General Purpose performance mode.

35. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
EFS distributed data storage design
File systems distributed across
unconstrained number of servers
• Avoids bottlenecks/constraints
of traditional file servers
• Enables high levels of
aggregate IOPS/throughput
Data also distributed
across Availability Zones
(durability, availability)
EC2
EC2
…
EC2
EC2
…
EC2
EC2
…

36. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
How to think about EFS relative to EBS
Amazon EFS Amazon EBS PIOPS
Performance
Per-operation
latency
Low, consistent Lowest, consistent
Throughput
scale
10+ GB per second About 2 GB per second
Characteristics
Data
availability/
durability
Stored redundantly across multiple
AZs
Stored redundantly in a single AZ
Access
1 to 1,000s of EC2 instances, from
multiple AZs, concurrently
Single EC2 instance in a single AZ
Use cases
Big data and analytics, media
processing workflows, content
management, web serving, home
directories
Boot volumes, transactional and
NoSQL databases, data
warehousing, and ETL

37. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
An implication of per-operation
latency: I/O size impacts throughput
of serialized operations
4 KB 32 KB 256 KB 2 MB 16 MB
I/O size
Throughput

38. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
How to take advantage of EFS’s
distributed architecture: Parallelize
0
5000
10000
15000
20000
25000
30000
0 20 40 60 80 100 120 140 160
IOPS
# of Total Threads

39. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Kernel version and mount options
Recommended
Kernel
version
• Use Linux kernel 4.0+ (e.g., Amazon Linux 17.09.0—use latest
versions, Ubuntu 16.04 or 17.10)
Mount
options
• Mount via NFSv4.1
• Specify 1MB read/write buffers (“rsize”/”wsize”)
• Ensure operations are asynchronous
Recommend the following mount options:
-o nfsvers=4.1, rsize=1048576, wsize=1048576,
hard, timeo=600, retrans=2, async

40. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Bursting model: Throughput you can
drive scales as your file system grows
50 MiB/s per TiB continuously
100 MiB/s per TiB in bursts

41. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Phase 3:
Ingest

42. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
What do you think about?
Where it’s coming from How to move it
as fast and easily
as possible

43. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Where is it coming from?
Corporate data
center
Amazon
S3
Amazon EBS Other on-cloud
file systems

44. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
AWS Direct Connect
AWS
Direct Connect
EFS in your Amazon VPCOn-premises servers

45. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Third-party VPN
Amazon
EFS
Mount targetVPN
instanceNFS client VPN

46. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
By the way, Amazon EFS over AWS DX or
third-party VPN isn’t just for migration
Bursting
Migration
• Move entire data set permanently to EFS
• Access the data from applications running on EC2 instances
• Move data set temporarily to EFS
• Access the data from applications running on EC2 instances
• Move data back on premises when processing finishes
Backup and
disaster recovery
• Maintain copy of entire data set on EFS
• Restore the data to on-premises storage or (for DR) access
the data from failed-over applications running on EC2
instances

47. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
How to leverage parallelism to copy
data faster
rsync cp fpsync mcp
cp +
GNU parallel
fpart + cpio +
GNU parallel
single-
threaded
single-
threaded
multi-
threaded
multi-
threaded
multi-
threaded
multi-
threaded
Poor
(very chatty)
Fair Good Better Better Best

48. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Results
11.7
15.6
24.4
58.9
70.1
93435
329
210
87
73
55
0
175
350
525
0
20
40
60
80
100
rsync cp fpsync mcp cp + GNU
Parallel
fpart + cpio +
GNU Parallel
Seconds
MB/s
File Transfer Tools
Throughput (MB/s) Duration (seconds)
EC2 instance:
c4.2xlarge
Transfer test:
5,000—1 MiB
files (5 GiB) from
EBS to EFS

49. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Introducing EFS File Sync
Fast
Up to 5x faster than
standard Linux copy tools
Sync data from existing file systems into Amazon EFS file systems
Secure
Encrypted parallel data
transfer to AWS
Simple
Set up and manage easily
from the AWS Console
file systems from on-premises to EFS
DIY in-cloud file systems to EFS
EFS file systems between AWS Regions
Use EFS File Sync to copy…

50. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Within your VPC
EFS File Sync
DIY with fpart + cpio +
GNU parallel
Outside your VPC
EFS File Sync
How to think about the options

51. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
EFS File Sync demo

52. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Between Amazon EFS & Amazon S3
Mount
target
NFS clientS3 bucket
Amazon
EFS

53. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Between Amazon EFS & Amazon S3
S3 bucket
Amazon
EFS
Mount
target
NFS
clients

54. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Phase 4:
Run

55. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
What do you think about? Not much…
View file system metrics
Amazon CloudWatch metrics
AWS CloudTrail API access
logs
Perform backups

56. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
CloudWatch metrics
DataReadIOBytes
DataWriteIOBytes
MetaDataIOBytes
TotalIOBytes
BurstCreditBalance
PermittedThroughput
ClientConnections
PercentIOLimit
Amazon
CloudWatch

57. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
CloudTrail API access logs
AWS
CloudTrail
Logs EFS API calls
from Console, SDK, CLI
Amazon S3
bucket

58. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Reference Architecture: WordPress

59. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Reference Architecture: Drupal

60. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Reference Architecture: Magento

61. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
EFS to EFS
Automatically backs up EFS
Easy to deploy
Backup solution

62. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Wrapping up

63. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
How to learn more: aws.amazon.com/efs
Feature blogs
White papers
Ref architectures
TCO calculator
10-minute tutorials
Documentation

64. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
For Enterprise Storage Engineers
• Learn how to architect and
manage highly available
solutions on AWS storage
services
• Advance toward AWS
certifications
• Help your organization migrate
to the cloud faster
Online at www.aws.training
• Access 100+ new digital
training courses including
advanced training on storage
• Deep Dives on S3, EFS, and EBS
• Migrating and Tiering Storage
to AWS (Hybrid Solutions)
At re:Invent
• Visit Hands-on Labs at the
Venetian
• Attend a proctored
“Introduction to EFS” Spotlight
Lab on Thursday at 3 pm at the
Venetian
• Meet Storage experts at the Ask
the Experts in Hands-on Labs
room at the Venetian
New Storage Training

65. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
re:Invent EFS Sessions
Monday, 11:30AM STG314: Case Study: Learn How HERE Uses JFrog Artifactory with Amazon EFS
Tuesday, 12:15PM STG401: This Is My Architecture - Lightning Round: Most Innovative Storage Solutions
Tuesday, 2:00PM STG327: Chalk Talk: Learn How to Maximize Throughput to EFS by Using Parallel Data Access Methods
Tuesday, 2:30PM STG324: Learn to Build a Cloud-Scale WordPress Site That Can Keep Up with Unpredictable Changes
Tuesday, 4:45PM STG307: Deep Dive on Amazon Elastic File System (Amazon EFS)
Wednesday, 12:15PM STG325: Case Study: Learn How SiriusXM & Digital ReLab Leveraged Amazon EFS for their Media Workflows
Wednesday, 4:00PM STG310: Files in AWS: Overcoming Storage Challenges for Common File Use Cases - Special Guest Celgene
Thursday, 12:15PM STG307-R: Deep Dive on Amazon Elastic File System (Amazon EFS) [REPEAT]
Thursday, 3:15PM STG315: Case Study: Learn How Thomson Reuters Uses Amazon EFS to Deliver Billions of Pieces of Content
Thursday, 3:15PM STG321: Amazon EFS: Leverage the Power of a Distributed Shared File System in the Cloud [WORKSHOP]
Thursday, 5:30PM STG326: How to Build a Scalable & Highly Available Messaging Service with IBM & TIBCO Using Amazon EFS

66. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Thank you!