Most likely, your organization is not in the business of running data centers, yet a significant amount of time and money is spent doing just that. Amazon Web Services provides a way to acquire and use infrastructure on-demand, so that you pay only for what you consume. This puts more money back into the business, so that you can innovate more, expand faster, and be better positioned to take advantage of new opportunities. Learn from the CEO of DevFactory on how they saved money and redirected their resources towards boosting innovation after taking advantage of the cloud.
2. Agenda
Why the Economics of Cloud is Different?2.
AWS Cloud Benefits Categories3.
Customer story : Dev factory4.
AWS Differentiation and Global Infrastructure1.
Turning the business case into reality5.
Introduction to AWS Cost Optimization6.
4. What sets AWS apart?
Building and managing cloud since 2006
90+ services to support any cloud workload; rapid
customer driven releases
16 regions, 44 availability zones, 91 edge locations
Thousands of partners; 3,800+ Marketplace products
Experience: 1M+ customers
Service Breadth & Depth; pace of innovation
Global Footprint
Ecosystem
Fine-grained controlSecurity
Fully integrated in AWSArtificial Intelligence
Gartner Magic quadrant recognizing AWS for
7th year
Enterprise leader
5. AWS Global Infrastructure
16 Regions – 44 Availability Zones – 91 Edge Locations
Region & Number of Availability Zones
AWS GovCloud (2) EU
Ireland (3)
US West Frankfurt (2)
Oregon (3) London (2)
Northern California (3)
Asia Pacific
US East Singapore (2)
N. Virginia (5), Ohio (3) Sydney (3), Tokyo (3),
Seoul (2), Mumbai (2)
Canada
Central (2) China
Beijing (2)
South America
São Paulo (3)
Announced Regions
Paris, Ningxia, Stockholm, Hong Kong
US govt cloud(US-East)
8. AWS Pricing Principles
Pay as you go
Pay less by using more
Pay less when you reserve
Pay less when AWS grows
No up-front investment
Pay per use
9. AWS Pricing Models
Pricing
Models
On-demand
Reserved
Instances
Spot
• Users that prefer the low
cost and flexibility of
Amazon EC2 without any
up-front payment or long-
term commitment
• Applications with short-
term, spiky, or
unpredictable workloads
that cannot be interrupted
• Applications being
developed or tested on
Amazon EC2 for the first
time
• Applications with steady
state usage
• Applications that may
require reserved capacity
• Customers that can
commit to using EC2
over a 1 or 3 year term to
reduce their total
computing costs
• Applications that have flexible
start and end times
• Applications that are only
feasible at very low compute
prices
• Users with urgent computing
needs for large amounts of
additional capacity
11. Benefit Categories of Cloud Computing
Operational CostsWorkforce ProductivityCost AvoidanceOperational ResilienceBusiness Agility
• Reduced Risk
Profile/Reduced
cost of Risk
Mitigation
• Revenue & margin
improvements due
to reduced outages
• Reduced Time to
Market
• Increased
operational agility
(new market
penetration,
divestiture,
acquisition)
• Reduced cost &
increased Pace of
Experimentation
• Elimination of
hardware refresh
programs
• Elimination of
maintenance
programs
• Unit Price of
Infrastructure
• Ability to match
supply & demand
• Reinvention is
continuous – a true
efficiency dividend
• A pathway to
optionality
• An elastic Cost Base
• Transparency drives
a lean mindset
• Automation drives
maintenance
efficiencies
• Reduced cost of
planned and
unplanned outages
• Increased
Developer
Productivity
12. GE Oil & Gas experienced business benefits
across all categories
13. Other Examples
RESOURCE EFFICIENCYCOST SAVINGS OPERATIONAL RESILIENCE BUSINESS AGILITY
80% reduction in software R&D times
(Apeejay Stya & Svrán)
Time to launch digital campaigns cut from
weeks to 24 hours (91App)
Calc and reporting time cut from 10 days to
10 minutes (Aon Benfield)
Time to market cut from weeks to hours
(FlyDubai)
Clinical simulations 98% faster than on-
premise (Bristol-Myers Squibb)
Time to deploy IT compute reduced to <
5minutes (Alcatel-Lucent)
R&D RFS times reduced from 6 months to 1
day (NewsCorp)
Provisioning time cut from 3-4 weeks to 2
days (ENEL)
Test-run time cut to 10 minutes, from up to
two hours (Yelp)
Launch new products 75% faster (Unilever)
5M financial investment policies recalculated
in minutes instead of overnight (ABSI)
DC footprint from 13 to 6 (The Weather Co.)
DC footprint reduced from 8 to 3 by 2018
(CapitalOne)
Storage costs reduced by up to 60% (ENEL)
Migrated 500 servers, 1PB storage, &100 DB
servers in 3 months to AWS (Condé Nast)
Over 50% reduction in TCO (GE)
DC footprint from 45 to 6 (News Corp)
50% reduction in app costs (Time Inc)
Cost savings of $20M p.a. (FINRA)
Computational cost reduced by 20%+ (ENEL)
Cloud deployment has saved US$34 million in
CAPEX and reduced OPEX by 85% (Samsung)
Queries run on 450,000 online subscribers 98
percent faster and 80 percent cheaper (FT)
Cost reduction of $40k p.a. (Dow Jones )
Savings of $1.5M p.a. (Trainline)
30% reduction in OPEX (MacMillan)
12% reduction in OPEX (RWE Czech)
Scaled to handle a 400% increase in page
views (Kurt Geiger)
Improved security posture (CapitalOne)
8600 transactions/second (McDonalds)
Transfer of over 750 TB of data from pipeline
inspection machinery (GE)
Processing over 75 billion market events daily
(FINRA)
Critical applications run in multiple AZs, x-
Regions for robust disaster recovery (Expedia)
Supports over 300,000 requests per minute to
its API (Easy Taxi)
60% reduced downtime (Trainline)
Migration of SAP on Oracle to AWS with zero
unplanned downtime across five countries
(Kellogg’s)
SAP availability boosted to 100% (MacMillan)
Over 200 Solutions Architects trained
in cloud (Hitachi)
Energy Marketing business prepared
for acquisition in only 6 months rather
than 12 (Hess Corp)
Performance targets over-achieved
by 43-66% (McDonalds)
IT Infra consolidation completed in
20% of expected time (Hearst)
25,000 Amazon Workspaces cloud-
based desktops deployed (Johnson &
Johnson)
Over 500 hours per year of server
configuration time saved (Sage)
39 years of Computational chemistry
condensed into 9 hours (Novartis)
15. AWS Cloud Benefits Categories
TCO
Resource Efficiency
Business
Value
Specificity
Accelerate Application
Deployment
Global Expansion
M&A or Divestures
Risk Mitigation
Business
Drivers
IT Drivers
Director of IT Executive Level
16. Driver #1:TCO - What is TCO and how do we do
it?
≠
Comparing TCO isn’t easy
Traditional Data Center
& Co-Location
17. Driver #1: TCO
Definition: Comparative total cost of ownership analysis (acquisition
and operating costs) for running an infrastructure environment end-to-end
on-premises vs. AWS.
Used for:
1) Comparing the costs of running an entire infrastructure environment or
specific workload on premises or in a co-location facility versus on AWS.
2)Budgeting and building the business case for moving to cloud
3) Paralleling an existing AWS workload with an on premises or co-location setup
18. How do customers lower their TCO with AWS?
Source: IDC Whitepaper, sponsored
by Amazon, “The Business Value of
Amazon Web Services Accelerates
Over Time.” December 2013
“Customers will have spent
63.4% more on average on-
prem or in co-location”
Remove over
provisioning and
move to pay for what
you use model
62 Price
Reductions
Economies of scale
allow AWS to
continually lower costs
Pricing model choice
to support variable &
stable workloads
Save more money as
you grow bigger
On-Demand
Reserved
Spot
Tiered Pricing
Volume Discounts
1 2 3 4
19. “…we don’t make money
when we sell things. We
make money when we help
customers make purchase
decisions.”
Jeff Bezos (HBR 2013)
20. Continually lowering prices for customers is in
our DNA
Reduced
Prices
More
Customers
More AWS
Usage
More
Infrastructure
Economies of
Scale
Lower
Infrastructure
Costs
Ecosystem
Global Footprint
New Features
New Services
Infrastructure
Innovation
We pass the savings along
to our customers in the
form of low prices and
continuous reductions
21. A typical on-premises compute environment is
massively underutilized
Used IT
Capacity
Idle
Capacity
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
On-Premises IT
Compute capacityTotal
Studies by Gartner, McKinsey and the
Uptime Institute have stated that typical
data centers are on average
less than 50% utilized
www.uptimeinstitute.org
anthesisgroup.com/wp-content/uploads/2014/08/Data-Center-Issue-Paper-final826.pdf
www.nytimes.com/2012/09/23/technology/data-centers-waste-vast-amounts-of-energy-belying-industry-image.html
22. Part of this can be explained by buying for “peak
load” requirements with inflexible infrastructure
Application/Workload drivers
Fluctuating/“Spiky” Part-time Cyclical
Peak
Peak
Peak
Used IT
Capacity
Idle
Capacity
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
On-Premises IT
Compute capacityTotal
23. Typical TCO considerations
Hardware – Server, Rack
Chassis PDUs, ToR
Switches (+Maintenance)
Software - OS,
Virtualization Licenses
(+Maintenance)
Facilities Cost
Hardware – Storage
Disks, SAN/FC Switches
Software - Backup
Network Hardware – LAN
Switches, Load Balancer
Bandwidth costs
Software – Network
Monitoring
Server Admin, Virtualization Admin, Storage Admin, Network Admin, Support Team
Diagram doesn’t include every cost item. E.g. software costs can include database, management, middle tier software costs.
Facilities cost can include costs associated with upgrades, maintenance, building security, taxes etc. IT labor costs can
include security admin and application admin costs.
Space Power Cooling
Project planning, Advisors, Legal, Contractors, Managed Services,
Training, Cost of capital
Business Value:
Cost of delays
Risk premium
Competitive abilities
Governance
Etc.
IT Labor Costs
Network Costs
Storage Costs
Server Costs
4
1
2
3
Extras5
Facilities Cost
Space Power Cooling
Facilities Cost
Space Power Cooling
24. Traditional Capacity Planning
Limited Value, Wasted $, Increasing Cost
Initial Fixed
Capacity
Utilization
Unused Capacity = Wasted $
Downtime, Lost Customers, Lost Revenue (Impossible to measure)
More Wasted $
Time
25. Initial questions to consider when exploring TCO
Capacity
Planning
1 How do you plan for capacity?
How many servers have you added in the past year? Anticipating next year?
Can you switch your hardware on and off and only pay for what is used?
Utilization
2 What is your average server utilization?
How much do you overprovision for peak load?
Operations
3 Will you run out of data center space some time in the future?
What was your last year power utility bill for the Data Center(s)?
Have you budgeted for both average and peak power requirements?
Optimization
4 Are you on AWS today?
Is your architecture cost-optimized (Auto Scaling, RIs, Spot, Instances turn
on/off)?
27. What levers are available to customers to drive down
costs?
Matching Supply and Demand – Right-sizing
Matching Supply and Demand – Elasticity
Lowering the Unit Price – Reserved and SPOT Instances
28. Example – An APAC Airline
Three Year TCO
Environment On-premises (US$) Lift & Shift (US$)
Right-sized on AWS
(US$)
AWS Optimized Apps
(US$)
Test 6,852,872 4,468,062 3,874,600 2,761,619
Dev 7,102,348 4,506,412 3,937,795 2,731,962
Prod 12,313,415 9,965,267 8,524,232 6,266,904
Total 26,268,635 18,939,741 16,336,627 11,760,485
Savings over
on-premises (%)
27.9% 37.8% 55.2%
Additional savings of
13.7% over lift and
shift approach
Additional savings of
28% over right-sized
approach
Scope: Servers acquisition and operational costs, storage acquisition and operational costs, networking costs and manpower.
29. We have identified additional AWS savings by
modeling both optimization and re-platforming
On-
Premises
Lift & Shift Instance Right-
Sizing
Improved
Elasticity
Storage
Optimization
Optimized Lift
and Shift
Measure
Monitor and
Improve
Serverless
Architecture
Managed
Services
Replatformed
AWS
Optimized
Traditional TCO Comparisons
30. Driver #1: TCO Examples
Three years on, we’ve saved
over $100 million in avoided
capital and are about 65% in
the cloud.
- Dominic Shine, CIO News
Corp1
(1) Geekwire, December 2016
We’ve been able to seamlessly
scale our infrastructure, better
serve our customers across
the globe, and reduce our fixed
costs by 75% and operational
costs by 83%.
- Valentino Volonghi, CTO,
AdRoll2
(2) AWS Public Case Study
We’ve realized a 52 percent
reduction in TCO. That stems
from a number of factors... [a
push for self-service, dynamic
storage, using lower cost VMs]
Ultimately these savings are a
byproduct of doing the right
thing.
– Ben Cabanas, CTO, GE
Transportation3
(3) AWS Blog, May 2016
31. Driver #2: Resource Efficiency
Definition: a comparison of staff time spent performing functions and
tasks in the legacy IT organization versus AWS
Common Measures:
1) Database Administrators eliminating time spent on DB patches
and upgrades
2) Storage Managers no longer required to test or “burn in”
additional object storage arrays before operating them
32. Driver #2: Resource Efficiency Examples
“[Our staff] spends 60% of work
on things like data proliferation,
lack of standards, security
hardening.. AWS is addressing
these needs…”
- Raji Arasu, Senior VP of Platform
and Services, Intuit1
(1) Geekwire, December 2016
Pharma Services
Industry Customer
51%efficiency improvement
in Compute related functions
20%improvement in
Application Development functions
33. Driver #3: Accelerate Application Deployment
Definition: Shortening the amount of time to deploy new applications &
features. (Hint: Think DevOps as a first step)
Common Measures: “time to deploy to production,” “deployment
frequency” or similar variants.
34. Driver #3: Accelerate Application Deployment
• Deployment cycle time reduced: 4 days
to hours
• +240% improvement in speed-to-market
• Customer functionality delivered 5x
faster
• Shortened it’s Build Phase by 90 daysPharmaceutical
Services Provider
35. Driver #4: Risk Mitigation
• Heightened security
• In the event of power failure
• Natural disaster
• Quick data recovery on AWS
• Greater administrative and security control using AWS IAM, cloud watch and cloud trail
• Greater security and performance improvements by using VPC and direct connect.
https://aws.amazon.com/solutions/case-studies/seaco/
36. Driver #5: Global Expansion
Definition: Improved capability to experiment rapidly with new products or to
expand in new regions/markets .
What’s the value?
• Rapid entry into a new market
• Lower cost of expansion
• Minimize exit costs
37. $(2,000,000)
$(1,000,000)
$-
$1,000,000
$2,000,000
$3,000,000
$4,000,000
$5,000,000
1 2 3 4 5
On-Prem v. AWS Global Expansion
AWS 5-year Forecast On-Prem 5-year Forecast
Driver #5: Global Expansion
• Expansion time down
reduced in half
• Eliminated $2.4M in
global expansion costs
• Recognized additional
$5M in revenue
Net financial impact of leveraging AWS for was $8.4M over base case
Automotive Services Provider
38. Driver #6: M&A or Divestitures
Definition: Accelerating the value of bringing on new businesses or shedding
of assets to accelerate a customers’ business goals.
What’s the value?
• Speed to market
• Faster transition time
• Lower transition costs
39. Driver #6: Divestiture Example
1. Accelerate focus on
upstream business.
2. Reduce transition timeline
from 12 to 6 months
3. Sold the business for $2.6b
to reinvest in upstream
business.
“We had an obligation to Hess to make
sure that infrastructure could be
handed off in short order, both legally
and physically. Using the cloud enabled
us to meet these requirements and we
are thrilled with the results.”
- Jim McDonald, Lead Architect
Customer Benefits Impact
40. • ESW Capital and the Trilogy Group – Transformed from an enterprise software
firm into private equity.
• Portfolio of 70+ Enterprise software companies
• DevFactory is the Transformation Factory for the group
• DevFactory has been working with AWS since 2007
AWS is one of the core pillars of our transformation runbook
41. We would move 100% of our infrastructure to AWS even if there weren’t any cost
benefits
42. • We’ve stopped spending time and resources on managing basic infrastructure.
• We’ve moved our investments up the stack in areas like
• Machine learning
• Advanced graph analytics
• Research on preventing bad code from getting into our codebases
Business Benefits
43. And by the way…
AWS allows us to bring our infrastructure costs down by over 90%
44. Lift and Shift Right Sizing
Database and
Storage
Consolidation
Docker
• Cost neutral
• Ensures that
products and apps
are portable
• Puts all
infrastructure under
standard
management
runbook
• Reduces costs by
40%
• Eliminates waste
• Gives us a true
picture of what is
really being used
• Aurora migration
• Additional 20%
cost reduction
• Helps improve
scale of the apps
• 90+% Cost Reduction
• Bigger is better – use
x1.32XLs
• Apps are at least 2x
more performant than
on dedicated
hardware.
Our 4 step Runbook
45. • Email backup and archiving solution - acquired from Dell in 2015
• $6M annual spend on data center
• Monolithic architecture, specialized Dell storage hardware
• Took 12 months to transform
• Moved to S3, MySQL Aurora and Docker
Current spend < $200K a year (96% reduction in COGS)
MessageOne Case Study
47. • The firewall for code is an active system that constantly monitors code that is getting checked in,
analysis it for potential problems and takes both preventative and corrective action:
• Builds
• Static Analysis and symbolic execution based analysis
• Test executions, mutation test runs
• Anti Pattern detection
• Automatic fixes to code
• Automatic code reviews
• The firewall for code cluster is an ECS Spot Cluster
• Runs a custom fleet optimization algorithm to ensure spread across all availability zones,
stability of the cluster and lowest costs.
• Has the ability to dynamically scale up or down depending on resource demands
Costs ~75% less than on-demand instances
Firewall for Code - Docker Spot Platform
Spot Instances
48. Turning the Business Case into Reality
Kaplan Closes 8 Data Centers.
By Migrating Data Centers, Wilmar Saves 50% in CAPEX.
Delaware North Moves Data Center to Save $3.5M
Over Five Years.
Hess Uses AWS to Streamline Data Center Migration
in 6 Months.
News UK Shortens Time to Market by 6 Months by
Migrating 60% of its Data Centers.
ENEL has migrated 10,000 servers and 6 PBs of data
to AWS in 9 months and saving up to 21% on
compute and up to 60% on storage costs.
53. The Five Pillars of Cost Optimization
Right-Sizing Your
Instances
Pick the Right
Pricing Model
Increase
Elasticity
Measuring &
Monitoring
Match usage to
storage class