Auto Scaling allows cloud resources to scale automatically in reaction to the dynamic needs of customers. This session shows how Auto Scaling offers an advantage to everyone—whether it's basic fleet management to keep instances healthy as an Amazon EC2 best practice, or dynamic scaling to manage extremes. We share examples of how Auto Scaling helps customers of all sizes and industries unlock use cases and value. We also discuss how Auto Scaling is evolving to scaling different types of elastic AWS resources beyond EC2 instances. Data Scientist & Principal Investigator, Hook Hua, from NASA Jet Propulsion Laboratory (JPL) / California Institute of Technology will share how Auto Scaling is used to scale science data processing of remote sensing data from earth-observing satellite missions, and reduce response times during hazard response events such as those from earthquakes, hurricanes, floods, and volcanoes. JPL will also discuss how they are integrating their science data systems with the AWS ecosystem to expand into NASA’s next two large-scale missions with remote-sensing radar-based observations. Learn how Auto Scaling is being used at a global scale – and beyond!

1. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
AWS re:INVENT
Auto Scaling
T h e F l e e t M a n a g e m e n t S o l u t i o n f o r P l a n e t E a r t h
A n d r é D u f o u r , S e n i o r M a n a g e r , A W S
H o o k H u a , D a t a S c i e n t i s t , J e t P r o p u l s i o n L a b o r a t o r y , C a l i f o r n i a
I n s t i t u t e o f T e c h n o l o g y
C M P 2 0 1
N o v e m b e r 2 9 , 2 0 1 7

2. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
These are a few of my favorite (and least
favorite) things

3. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Availability with fleet management
Keeping up with demand with dynamic scaling
Future of Auto Scaling

4. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Myth Fact
My application does not scale,
so I don’t need Auto Scaling
It’s hard to use
My instances are stateful or unique;
I can’t use Auto Scaling
It works well
with stateful instances
You can get started in minutes
It monitors and heals instances

5. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Auto Scaling groupAuto Scaling group
Auto Scaling
Fleet management Dynamic scaling
ELB
EC2 instances
ELB
CPU
Utilization
EC2 instances

6. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Is fleet management for you?
“I’ve got instances serving a
business-impacting application”
“If my instances become unhealthy,
I’d like them replaced automatically”
“I would like my instances distributed
to maximize resilience”

7. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Demo

8. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Dynamic scaling
Demo & Details: CMP311 “Auto Scaling Prime Time:
Target Tracking Hits the Bullseye at Netflix”
Thursday 12:15 p.m. - Venetian

9. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Scaling on a schedule
Recurring scaling
events
Schedule individual
events
Auto Scaling group
ELB
EC2 instances

10. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Reactive scaling with target tracking
1. Pick your
metric
2. Set the
target value
4. Profit
…
You do this part Auto Scaling manages your capacity

11. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Dynamic scaling for fun and profit
Collect metrics
Auto Scaling
Scaling policy:
keep my CPU at 50%
ELB
Auto Scaling group
50%50%65%
Add capacity
50%

12. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
NASA JPL: Auto
Scaling Advantage
Advanced Rapid Imaging and Analysis
© 2017 California Institute of Technology. Government sponsorship acknowledged.
Reference herein to any specific commercial product, or service by trade name, trademark, manufacturer, or otherwise, does not constitute or
imply its endorsement by the United States Government or the Jet Propulsion Laboratory, California Institute of Technology.

13. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Challenge: Automated and rapid remote
sensing for urgent disaster response
Automated data system are required to analyze large
quantities of data from NASA NISAR, other satellite missions,
and rapidly expanding GPS networks
Mean Access Time (Day)
∞ 4 2 1.3 1
Going from artisan to automation: Use system
engineering approach to translate specialized data
analysis into operational capability
Orbit Data Coherence
Library
AZO / MAI
EQ
Kinema c
Modeling
Level-0 Data
Kinema c
Model
Catalog &
Waveform
Sta c
Model
InSAR Processing
Time Series
Con guous Pairs Sta c
Modeling
Int. Phase
Stress Change
III. Rou ne Data Processing
Coherence
Change Maps
Seismometer
Radar Sensor
I. Geode c
Sensors
II. Data Providers / Archives
Global CMT
USGS NEIC
IRIS (Waveform)
Fault
Geometry
Mw>5.5 Mw>6.5
Yes
Yes
No Small EQ
Modeling
Amp / Arr. T
correc on
IV. Triggered
Legend
Ready to be used ( > 70 % readiness )
Need modifica on ( 30 – 70 % readiness)
Doesn’t exist ( < 30 % readiness) Latency / Bo leneck
Out of scope of R&TD
Customer Products
Level-0 Data
Op cal Im.
Processing
Offset Image
Event Driven
Op cal Sensor
SPOT Image (SPOT 1-5,
Formosat 2)
NSPO (SPOT)
DigitalGlobe (Quickbird,
Worldview 1)
RINEX Data
GPS Processing
Daily
Point Pos.
1 Hz (GDGPS)
Hourly
Daily
GPS Sta on
Orbit Data
Sub-daily
Point Pos.
SOPAC
+ many others
GPS Constella on
NASA - JPL
V. Customer
Products
Damage
Assessment
3D Surface
Disp. Maps
Ground
Mo on
3D Surface
Disp. Maps
EQ catalog
CTBT
correc on
surface
Coseismic
Interferogram
Coseismic GPS
vector map
Coseismic
Interferogram
Coseismic GPS
vector map
Temporal records
of ground deformation
Spatial maps of ground
deformation
Earthquake
models
Coseismic ground
deformation
Demonstrate response to hazards with standardized set of data products for decision and policy
makers
Coseismic damage

14. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Automated
data collection
& processing
Radar
Sensors
Building damage
and inundation
Radar
GPS Seismology
Permanent
ground deformation
High-resolution
hazard assessment
from fault models
Monitoring &
near real-time
assessment
Examples from the 2011 M9.0
Tohoku-Oki (Japan) earthquake
GPS
Networks
Seismic
Networks
Optical
Sensors
Advanced Rapid Imaging and Analysis
(ARIA)

15. Amatrice, Italy earthquake
(August 23, 2016)
Automated Urgent
Response Interferogram
Urgent response
processing of ESA’s
Sentinel-1A data to
interferograms were
automatically
processed—all in
AWS

16. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

17. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Hurricane Harvey Response
August 2017
Flood Proxy Map
Hurricane Maria Response
September 2017
Damage Proxy Map

18. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Urgent Response Analysis in AWS Cloud
M7.1 Earthquake near Puebla, Mexico (September 9, 2017)
Reactive Auto Scaling of
satellite analysis based on
increased data acquisitions

19. Urgent Response Analysis in AWS Cloud
SAR-based Damage Proxy Map

20. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Dynamic Scaling in Earth Science Data
System
The size of the science data system compute nodes can automatically
grow/shrink based on processing demand
Auto Scaling group policies
Target tracking scaling policies
Auto Scaling enabling runs
of over 100,000 vCPUs

21. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Earth Science Data System in AWS
NASA OCO-2 L2 full physics processing operational in AWS
• Processing of L2 full physics data products in Amazon cloud across multiple regions
• Scaled up thousands of compute nodes
• Demonstrated capability of higher internal data throughput rates than NISAR needs
Number of
compute
nodes over
time
Per node
transfer rate
over time
Scalable internal
data throughput
@ 32,000 full-physics
processing on 1,000 nodes
ASG max set to 1000 instances x 32 vCPUs

22. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Considerations for Scaling In/Out Events
• Target tracking scaling policies
• Scaling up in batches + rest periods
Scaling up (scale out)
• What policy to set to scale down?
CPU/network utilization
• Potential stateful domain knowledge only
known within the instances
• Instance protection
Scaling down (scale in)

23. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Auto Scaling and the Amazon EC2
Spot market
• Auto Scaling works well with Spot Instances
• Major cost savings (75%–90% savings over on-demand)…if you can use Spot Instances
• Compute instances terminated if market prices exceed your bid threshold

24. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Fleet Management for High Resiliency
X
X
X X
X
Availability Zone a Availability Zone b Availability Zone c
Running in Spot market forces the data
system to be more resilient to failures
Compute fleet instances

25. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Auto Scaling and the “Market Maker”
This OCO-2 data production run
of 1000 x 32vCPUs affected the
market prices
Strategy:
• Mitigate impact on spot market
• Diversification of resources
• “Spot fleet”

26. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
“Thundering Herd”
Fleet of ASG compute instances calling
same services at same time
• “API rate limit exceeded”
“Jittering” the API calls
• Introduce randomizations to API calls
• Distributes load on infrastructure
Service
Compute fleet instances

27. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Next generation NASA missions
• The volume of data
produced is larger
than previous missions
• Data storage,
processing,
movement, and costs
are the biggest
challenges
0
10
20
30
40
50
60
70
80
90
100
OCO-2 SMAP NISAR
Estimated Daily “Keep Up” Volume (TB)
(2021)(2009) (2015)

28. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
The future of Auto Scaling:
Application Auto Scaling

29. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Application Auto Scaling: Databases
Amazon
Aurora
Amazon
DynamoDB

30. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Application Auto Scaling: Compute
EMR Cluster
Task
Task
Task
ECS Service EC2 Spot Fleet

31. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Learn more
CMP301
Auto Scaling Prime Time: Target Tracking Hits the Bullseye at Netflix
Thursday, November 30, 12:15 p.m. - Venetian, Level 5, Palazzo P
aws.amazon.com/autoscaling/getting-started

32. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Thank you!