Cloud Economics: Optimising for Cost
•
4 likes•6,744 views
The Cloud provides a transparent cost model, with options from on-demand provisioning, to reserved capacity to low cost, spot pricing. This session outlines how to use these pricing models in concert to optimise the running cost of your application on the AWS Cloud. We'll outline new utilisation based reserved instances and bidding strategies when running on spot instances, to cost prediction.
Report
Share
Cloud Economics: Optimising for Cost
- 1. COST OPTIMIZATION Glen Robinson– Solutions Architect
- 3. Elasticity is a fundamental property of the cloud that drives many of its economic benefits
- 4. When you turn off your cloud resources, you actually stop paying for them
- 5. 6 am
- 6. 10 am
- 7. 25% Savings Web Servers 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Hour during a day
- 8. www.MyWebSite.com (dynamic data) Amazon Route 53 media.MyWebSite.com (DNS) (static data) Elastic Load Balancer Amazon Auto Scaling group : Web Tier CloudFront Amazon EC2 Auto Scaling group : App Tier Amazon RDS Amazon S3 Amazon Availability Zone #1 RDS Availability Zone #2
- 9. Auto scaling : Types of Scaling Scaling by Schedule Use Scheduled Actions in Auto Scaling Service • Date • Time • Min and Max of Auto Scaling Group Size You can create up to 125 actions, scheduled up to 31 days into the future, for each of your auto scaling groups. This gives you the ability to scale up to four times a day for a month. Scaling by Policy Scaling up Policy - Double the group size Scaling down Policy - Decrement by 1
- 10. Horizontal Scaling vs. Vertical Scaling
- 11. 75% Savings Web Servers 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Days of the Month during a month
- 12. End of the month processing Expand the cluster at the end of the month Expand/Shrink feature in Amazon Elastic MapReduce Vertically Scale up at the end of the month Modify-DB-Instance (in Amazon RDS) (or a New RDS DB Instance ) CloudFormation Script (in Amazon EC2)
- 14. Reserved Instances New Reserved instances Light Utilization Medium Utilization Heavy Utilization
- 15. Light Utilization RI 15-40% utilization Lower costs up to 33% Use Cases: Disaster Recovery, periodic processing (e.g. EMR)
- 16. Medium Utilization RI The old RI 41-79% utilization Lower costs up to 49% Use Cases: Web applications, many heavy processing tasks
- 17. Heavy Utilization RI > 80% utilization Lower costs up to 59% Use Cases: Databases, Large Scale HPC, Always-on infrastructure, Baseline
- 18. Optimizing Cost with RIs $18,000 $16,000 $14,000 $12,000 $10,000 Heavy Medium $8,000 Light O-Demand $6,000 $4,000 $2,000 $-
- 19. Optimizing Cost with RIs 14 12 10 On Demand 8 Light Utilization RI Medium Utilization RI 6 Heavy utilization RI 4 2 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
- 21. Instance Types
- 24. Basic recommendations on Instance Type Choose the EC2 instance type that best matches the resources required by the application Start with memory requirements and architecture type (32bit or 64- bit) Then choose the closest number of virtual cores required Scaling across AZs Smaller sizes give more granularity for deploying to multiple AZs
- 25. Tip – Instance Optimizer Free Memory Free CPU PUT 2 weeks Free HDD At 1-min intervals Alarm Amazon CloudWatch Instance Custom Metrics
- 26. Spot Instances Bid on unused Amazon EC2 capacity. Run instances as long as bid exceeds the current Spot Price
- 27. What are Spot Instances? Sold at Sold at 50% Unused 54% Unused Discount! Discount! Sold at Sold at 56% Unused 59% Unused Discount! Discount! Sold at Sold at 66% Unused 63% Unused Discount! Discount! Availability Zone Availability Zone Region
- 28. What is the tradeoff? Unused Unused Unused Reclaimed Unused Unused Reclaimed Unused Availability Zone Availability Zone Region
- 32. Strategy: Optimize for Cost Engineered application towards a cost Set low maximum bid price to minimize costs Were comfortable if process ran longer or jobs were re-run Did not pay for hour if they are interrupted
- 33. $$$$$$$$$$$$$$$$$$ $$$ $ $ $ $
- 34. Strategy: Price History Momentum Price Set 10% above Average Price Last Hour Maximum price threshold of 80% of On-Demand Price One time spot requests; one instance per request; across all availability zones Not more than 10 open Spot requests at any time Spot requests expire in 10 minute
- 35. Strategy: Price History Momentum
- 36. Strategy: Discount over On-Demand Bid around the On-Demand price Use On-Demand instance when Spot Price exceeds On-Demand price (or slightly higher) May pay more some hours, but on average they pay significantly less This bidding strategy ensures a discount over On-Demand
- 38. Strategy: Discount over On-Demand Switch to Substantial On-DemandBack Savings Switch to Spot Much lower costs than On-Demand
- 39. Strategy: Minimize Interruption ~57% Savings On Average Substantial Savings
- 43. $0.028 per hour DNS Elastic Load Web Servers Balancer Availability Zone $0.095 per hour (small instance) EC2 instance DNS + software LB Web Servers Availability Zone
- 44. Consumers Producer SQS queue $0.01 per 10,000 Requests ($0.000001 per Request) $0.095 per hour (small instance) Producer EC2 instance Consumers + software queue
- 45. Software v/s Services Software on EC2 SNS, SQS, SES Pros Pros • Custom features • Pay as you go • Scalability Cons • Availability • Requires an instance • High performance • SPOF • Limited to one AZ • DIY administration
- 46. 1. Elastic Capacity 2. Reserved Instances 3. Spot Instances 4. Instance Types 5. Complementary services