Amazon Web Services provides users with detailed utilization metrics, an unparalleled choice of resource sizes and types, multiple purchasing options, and advanced tagging capabilities. The question is how do you best leverage the rich data into actionable information?
Join CloudCheckr CTO Aaron Newman and AWS Solutions Architect Scott Ward to learn how a customer who is a leading image content company has controlled spend and satisfied finance across a complex environment with multiple consolidated billing accounts. This company was even able to identify $38k in monthly savings just by upgrading to optimal Amazon EC2 instance types across their 31 accounts!
The discussion will focus upon implementing cost management strategies such as:
• Purchasing Amazon EC2 Reserved Instances (RIs) and other reserved resources
• Using tags to track and allocate spend by business unit, department, or team
• Identifying orphaned, idle, and underutilized resources
• Creating budget alerts to ensure budget and cost discipline
5. Pricing Models on AWS
On-Demand
Pay for compute
capacity by the hour
For spiky workloads,
or to define needs
Reserved
Use No Upfront, Partial,
or All Upfront payment
options to receive a
significant discount on
the hourly charge
For committed
utilization
Spot
Bid for unused capacity,
charged at a Spot Price
which fluctuates based
on supply and demand
For time-insensitive or
transient workloads
Dedicated
Launch instances within
Amazon VPC that run
on hardware dedicated
to a single customer
For highly sensitive or
compliance related
workloads
Free Tier
Get Started on AWS
with free usage & no
commitment
For POCs and
getting started
6. Why Reserved Instances?
Scaling out
- Ensure that capacity you need to support your applications is there when you need
to scale out.
Failing over
- Ensure that the capacity you need to support your applications is there when you
need to fail over from your primary architecture.
Steady state
- You have a good understanding on your long term compute needs to commit to
the appropriate resources in exchange for reduced cost.
7. Reserved Instance Payment Options
No Upfront – 1 year term only
- Obtain reserved instance capacity with no upfront payments
- Discounted hourly rate for duration of term
- Up to 50% off compared to on-demand
Partial Upfront – 1 or 3 year term
- Balances payments between upfront and hourly
- Pay some money upfront and then pay low hourly rate for the rest of the term
- Up to 74% off compared to on-demand
All Upfront – 1 or 3 year term
- Pay all up front and no hourly fees
- Up to 75% off compared to on-demand
8. Reserved Instance Benefits
Capacity is reserved
- Confidence that instances can be launched when you need them
Discounts over on-demand workloads
- Up to 75% over on-demand
- Greater discounts on Reserved Instance Marketplace
Flexibility to modify Reserved Instance configuration based on current needs
Share pricing benefits, but not capacity reservation, among linked accounts
Resell unused/unwanted capacity in Reserved Instance Marketplace
10. What is resource tagging?
• Ability to add custom metadata to your AWS resources
• Allows you to categorize your AWS resources
– Purpose
– Owner
– Environment
– Project
– Cost Center
– Etc
11. Why is tagging important?
• Tool to organize resources for your customers
– Especially if you have customers with multiple linked accounts
• Run reports on specific tags across multiple accounts
• Use tags to control access to AWS resources
• Track project resources for a customer
• Track utilization across cost center
• Enhanced search capabilities
• Greater level of information in the detailed billing report
12. Uses for Tagging
• Searching for resources in the Management Console and APIs
• Grouping matching tags across AWS resources into Resource
Groups
– Allows you to view details about all resources at once.
– Track resources tied to a particular project
• Detailed billing information
– Detailed Billing Report
– Cost Explorer
• Controlling access policies based on tags
– Enhance IAM access policies to be fine grained
14. Billing Alerts
• Allow you to be notified when your estimated monthly spend
reaches a certain amount.
• Define alerts for your overall estimated charges or for specific
services.
15.
16.
17. AWS at Leading Image Content Company
Current
• In use for many years, but with no centralized standards, insight or focus (many
accounts came from different subsidiaries and/or companies we acquired)
• Some very large projects
• ~30 accounts to date
Future
• New projects appear about once or twice a month
18. Implementing CloudCheckr
History
• CloudCheckr brought in for evaluation in summer ’14
• Became corporate standard in Dec ‘14
• Account collection into AWS Master Billing
• CloudCheckr began in November ‘14 through April ’15
Focus
• Reduce cost
• Increase visibility
• Centralize and standardize analytics
19. CloudCheckr Strategy (so far…)
We’ll discuss each of these in the next few slides
• Understand and reduce cost and utilization across enterprise
• Leverage Best Practice reports to optimize usage
• Support charge-backs within projects and customer base
20. Identify Cost/Usage Across the Enterprise
• Identified all AWS accounts
Moved to Consolidated Billing and monitor using CloudCheckr
• Understand overall spend
Who’s spending what and where in one report
• Understand usage
Number of resources used
Identify defunct resources and accounts
What’s stored in S3 today
21. Utilize Best Practices to Optimize Usage
Best Practice Cost
• Identified $38k/year in immediate cost savings (new generation
instances)
• Identified another potential $20k/year in potential savings
(under-utilized instances)
• Identified ~$10k/year in a few accounts of accounts that were
deprecated (they were cancelled)
22. Migrate Previous Generation Instances
How much can you save?
http://aws.amazon.com/ec2/pricing/
http://aws.amazon.com/ec2/previous-generation/
Examples:
m1.large : vCPU:2 / ECU:4 / Memory(GiB) 7.5 / 2 x 420 $0.175 per Hour
m3.large : vCPU:2 / ECU:6.5 / Memory(GiB) 7.5 / 1 x 32 SSD $0.133 per Hour
24% savings - Do you need the extra storage? What about m4?
23. Support Charge-backs Within Projects
AWS accounts that "host" small projects that don’t want to be bothered
by infrastructure costs and support
• Using tags on AWS resources
• Run custom cost reports to charge back costs based on tags
AWS accounts that support paid customer access (in progress)
• Using tags on AWS resources
• Manually invoicing right now after running a report similar to
above
• Planning to leverage CloudCheckr invoicing
27. Building a Tagging Strategy
Determine the Tags you want
- E.g. Department, Owner, ProductLine, Expiration
Define your Tagging Rules
- https://app.cloudcheckr.com/#Report/ImproperlyTaggedResources
Monitoring for Improperly Tagged Resources
- Check daily or push emails
- Terminate instances that violate your policy
28.
29.
30.
31. Creating Advanced Budget Alerts
- Enhances the Billing Alarms in CloudWatch
- Build budget alerts from tags, project times, multiple thresholds, etc…
32. Best Practice Checks
350+ Best Practice Checks around Cost, Security, Availability, Usage
- Complete integration with Trusted Advisor
34. Drilling down into more details
Best practice checks drilldown into complete details, history, etc…
35. Rebalancing Reserved Instances
Adjust your Reserved Instances to best utilize what you have running
Modify the instance type and Availability Zone from the Console
- http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ri-modifying.html
36. Thank You for Attending
Sign up today for free evaluation
at http://cloudcheckr.com
Aaron Newman is the Founder
of CloudCheckr (www.cloudcheckr.com)
Please contact me with additional questions at:
aaron.newman@cloudcheckr.com
Notes de l'éditeur
Start out
Scott Ward is a solutions architect with Amazon Web Services (AWS). Based out of Seattle, WA, Scott is part of a team that supports AWS's global network of technology and consulting partners. For AWS partners Scott’s focus is on existing AWS architecture validation and improvements, planning new architectures for workloads on AWS, providing guidance on how to use new AWS services, and providing general technical enablement. Scott has a deep background in supporting, enhancing, and building global financial solutions to meet the needs of large companies, including many, many years supporting the global financial systems for Amazon.com
X X X
AWS has a variety of purchase options that allow you to match your workload to the right model, and can help you optimize your bill by working with you to choose the right mix of several of these. Let’s look at three models here; on-demand, reserved, and spot instances.
With on-demand instances, customers pay for compute capacity by the hour with no minimum commitments required. These instances free you from the costs and complexities of planning, purchasing, and maintaining hardware and transforms what are commonly large fixed costs into much smaller variable costs.
Reserved instances give you the option to make a one-time payment for each instance you want to reserve. This provides you with a significant discount (up to 75%) compared to On-Demand Instance pricing. In addition, Reserved Instances provide you with a capacity reservation, so you can have confidence that you will be able to launch the instances you have reserved when you need them.
With spot instances, you can bid for unused Amazon EC2 capacity. Instances are charged the Spot Price, which is set by Amazon EC2 and fluctuates periodically depending on the supply of and demand for Spot Instance capacity. To use Spot Instances, you place a Spot Instance request, specifying the instance type, the Availability Zone desired, the number of Spot Instances you want to run, and the maximum price you are willing to pay per instance hour.
On-Demand allows you to start and stop EC2 instances and you only pay for the time that the instances are running. This is great for the initial experimentation phase as well as systems that do not need high availability and may just be there to provide business support on on an occasional basis, or are a DR solution.
Reserved instances offer you the lowest hourly rates for a one time up front fee. Reserved instances are a way to drive value to our customers who have more of a steady state workload on their instances. They are available for 1 or three year terms and have pricing models to meet Light, Medium, or Heavy utilization workloads.
With Spot instances there is an opportunity to take advantage of un-used capacity in AWS. Since we have a large amount of infrastructure at AWS it stands to reason that at any point in time there are some boxes sitting idle waiting for customers. The spot market lets users take advantage of market-set pricing on those instances, which often results in an incredibly deep discount. Two points to keep in mind for spot instances:
The market constantly fluctuates with no guarantee of capacity availability or price.
When someone is willing to pay more for your instance than you are, we will abruptly terminate your instance and sell it to them. You are not locked in for any specific duration.
4) Dedicated instances are instances available within an AWS VPC for customers having workloads that require that their application run on hardware dedicated to a single customer. This is a good use case for highly sensitive or compliance related workloads.
Reserved instance types
-All Upfront – pay all up front and then get the lowest hourly rate.
-Partial Upfront – Pay partial upfront fee and then pay the rest as a regular monthly amount for each instance.
-No upfront – Pay nothing up front and pay a regular monthly fee for each instance for the duration of the RI term. Great way to get into using Ris without having to commit all up front right away.
These different payment options are here to help you make a purchase decision based on what you are able to support from a financial perspective.
You can do this for any service that you are using in your AWS account