Regardless of whether you do nothing, build kit, buy from AWS or another CSP, someone from finance will come back to you and ask what happened to their money. In this session we will cover Cloud ROI: the key economical drivers for moving to the cloud and the tips and tricks for cost optimization on AWS.
2. What to expect
We will introduce our framework for
optimizing your AWS bill by focusing
on paying only for what you need.
And leave you with actionable ideas
to take back to the office to
implement immediately.
7. But how do you
ensure that you
only pay for what
you need?
8. What do you need and how do you build for it?
Initial Fixed
Capacity
Utilization
Unused Capacity = Wasted $
Downtime, Lost Customers, Lost Revenue (Impossible to measure)
More Wasted $
Time
13. Best Practice Framework for Architecting on AWS
Performance
Ensure a system delivers maximum
performance for a set of resources.
Cost Optimization
Achieve the lowest price for a workload
taking into account fluctuating needs.
Reliability
Ensuring a given system is architected to
meet operational thresholds during a
specific period of time.
Security
Review definitions and compliance best
practices.
16. The Five Pillars of Cost Optimization
Right-Sizing Your
Instances
Pick the Right
Pricing Model
Increase
Elasticity
Measuring &
Monitoring
Match usage to
storage class
17. Pillar 1: Right-Sizing
Right-sizing
• Selecting the cheapest instance available
while meeting performance requirements
• Looking at CPU, RAM, storage, and network
utilization to identify potential instances that
can be downsized
• Leveraging Amazon CloudWatch metrics and
setting up custom RAM metrics
Rule of thumb: Right size, then reserve.
(But if you’re in a pinch, reserve first.)
18. Pillar 2: Increase Elasticity
Turn off nonproduction instances
• Look for dev/test, nonproduction instances that
are running always-on and turn them off.
• Lambda + CloudWatch = Automated Scheduling*
Autoscale production
• Use Auto Scaling to scale up and down based on
demand and usage (for example, spikes).
* https://aws.amazon.com/premiumsupport/knowledge-center/start-stop-lambda-cloudwatch/
20. Reserved Instances for Always-On Instances
Commitment level
• 1 year
• 3 year
AWS services offering RIs
• Amazon EC2
• Amazon RDS
• Amazon DynamoDB
• Amazon Redshift
• Amazon ElastiCache
* Dependent on specific AWS service, size/type, and region
21. New Reserved Instance Regional Benefit
The Reserved Instance Regional
Benefit allows customers to run
instances in any AZ in an AWS
Region and AWS will apply the
reservation pricing benefit to your
instances automatically.
22. Consider Convertible RIs where flexibility is required
With a Convertible Reserved Instance, you can modify
your existing reservation across:
Instance families
Instance sizes
Operating Systems
Tenancy
25. Reserved Instances
Step 1: RI Coverage
• Cover always-on resources with standard or
convertible RIs
Step 2: Increase RI Utilization
• Known architectures: Leverage Standard RI
flexibility to increase utilization.
• Growing or changing architectures: Leverage
Convertible RIs across families, sizes, and OS.
• Regional Benefit: Consolidated billing,
reservation not critical
26. Options
– Spot Fleet to maintain instance
availability
– Spot Block durations (1-6 hours)
for workloads that must run
continuously
Commitment level
– None
* Compared to On Demand price based on specific EC2 instance type, region, and Availability Zone
Consider Spot for Elastic Workloads
27. Consider Spot for Elastic Workloads
• Be Fault Tolerant
• Workloads should be Stateless
• Loosely Coupled workloads preferred
• If possible, deploy to Multiple AZs
• Instance Flexibility is king
• Take advantage of the 2 minutes warning
• There is always Spot capacity available
28. Spot Rules
50% of OD
75% of OD
25% of OD
You pay the market price
87% discount!
You’ll Never Pay More Than Your Bid
29. Novartis, Cycle Computing &
AWS Spot instances
Goal: Screen 10 million compounds against a
common cancer target
Estimated on-prem costs: $40M
33. Object Storage Classes on Amazon S3
Active data Archive dataInfrequently accessed data
Standard
Hot
Standard - Infrequent Access
Warm
Amazon Glacier
Cold
34. Standard-Infrequent Access (SIA)
Comparing S3-IA to S3 Standard
What’s the Same?
Access:
Almost identical to S3 standard with
same buckets and URLs
Durability:
11 “9’s” of durability
Performance:
Same characteristics as S3
Standard, low latency, high
throughput.
What’s Different?
Availability:
Designed for 99.9% availability, vs.
99.99% availability of S3.
Cost:
Lower Capacity fees, plus retrieval
fees. Minimum object size and
retention.
35. Running the Numbers: S3 or S3-IA
Comparing 1 PB of object storage*
S3 S3-IA Savings %
Rule of thumb: Breakeven ≈ 100% Retrieved per Month
Content Accessed
per Month
1PB Monthly $24,117 $14,116 41%10%
$24,117 $18,350 24%50%1PB Monthly
$24,117 $23,593 2%100%1PB Monthly
* Based on US-East Prices
41. Enabling Cost Transparency
Cost Allocation
Tagging
Must Have Tagging:
• Cost Center
• Application or Workload
• User
• Expiration Date
• Automation Support
http://ape.gs/AWSAutoTag
48. AWS Trusted Advisor
Helping Customers automate best practices (checks) across
cost optimization, security, fault tolerance, and performance
improvement
Red (action recommended)
Orange (investigation recommended)
Green (no problem detected)
54. Questions your Center of Excellence Should be Asking
1. How much of our workloads are “steady state”?
2. How are we currently handling our elasticity
needs?
3. Have we had a Well Architected review with AWS?
4. What’s keeping us from reserving capacity?
5. How can I be more involved in our process?
55. • Aligning Incentives (Carrots & Sticks)
• Automation
• Reporting
• Control & Governance
• Metrics / KPIs
Cloud Center of Excellence
57. Metrics and Targets
• % instances turned off daily
• % of instances right-sized
• % always-on resources covered by RIs
• % RI utilization
Define success by establishing metrics and track progress
Cost-Based
58. Value Based Optimization Metrics
• A company’s overall AWS cost should be evaluated as a unit cost ratio with
respect to another defined metric:
• 𝑈𝑛𝑖𝑡 𝐶𝑜𝑠𝑡 =
𝑇𝑜𝑡𝑎𝑙 𝐶𝑜𝑠𝑡
𝐼𝑛𝑑𝑖𝑣𝑖𝑑𝑢𝑎𝑙 𝑜𝑟 𝐵𝑢𝑠𝑖𝑛𝑒𝑠𝑠 𝑀𝑒𝑡𝑟𝑖𝑐
Examples
• Unit cost per customer or active subscriber
• Unit cost per revenue generated
• Unit cost per product or business unit
• Unit cost per internal user
• Unit cost per experiment
• Unit cost per FTE
Align to
Value Drivers