12 Ways to Manage Cloud Costs and Optimize Cloud Spend
0 likes•2,234 views
The document outlines 12 effective strategies for managing and optimizing cloud costs, highlighting issues such as wasted cloud spending and the importance of utilizing cost-saving features across cloud services. Key recommendations include using preemptible VMs, tracking resource utilization, automating workload management, and implementing tagging for better resource allocation. Additionally, the document suggests leveraging different pricing options, optimizing instance types, and monitoring cloud expenditures to achieve significant savings.
12 Ways to Manage Cloud Costs and Optimize Cloud Spend
- 1. 12 WAYS TO MANAGE CLOUD COSTS AND OPTIMIZE CLOUD SPEND
- 2. Panelists • Kim Weins • VP Marketing, RightScale
- 3. 30% 15% % of Cloud Spend Wasted Cloud Users Underestimate Wasted Spend Source: RightScale 2017 State of the Cloud Report Self-Estimated Wasted Spend Additional Wasted Spend Measured by RightScale
- 4. 3% 16% 19% 20% 23% 33% 30% 30% 31% 45% 4% 15% 20% 24% 25% 31% 33% 35% 38% 52% Use Google Preemptible VMs Use AWS Spot Instances Move workloads to cheaper cloud/region Select cloud or region based on cost Track AWS RIs to make sure they are used Purchasing AWS RIs Look for storage volumes not in active use Shut down workloads during certain hours Automate shut down of temporary workloads Monitor utilization and rightsize instances How Companies are Optimizing Cloud Costs 2017 2016 Companies Increase Focus on Cloud Costs Source: RightScale 2017 State of the Cloud Report
- 5. Two Solutions from RightScale VIRTUAL SERVERS PUBLIC CLOUDS IAAS+/PAAS SERVICES PRIVATE CLOUDS BARE METAL SERVERS CONTAINER CLUSTERS MULTI-CLOUD ORCHESTRATION AND GOVERNANCE RIGHTSCALE OPTIMA Collaborate across cloud governance teams, business units, and resource owners to manage and optimize cloud spend RIGHTSCALE CMP Orchestrate, automate, and govern applications across any cloud, any cloud service, any server, and any container.
- 6. RightScale Optima: Key Capabilities 5 Cost Aggregation Analysis Allocation & Reporting Forecasting Budget Controls Provisioning Policies Recommendations Collaboration Dynamic Policies
- 7. Cost Review Example 30-45% Savings 6
- 8. • Tagging • Cost Allocation • Reporting • Forecasting • Budget Enforcement • Provisioning Policies Manage Cloud Costs Optimize Spend • Underutilized Resources • Unused/Idle Resources • Dev Environments • Regional Cost Differentials • Superseded Instance Types • Discounting Options 12 Ways 7
- 10. Tagging: Define Required Global Tags Tag Type Examples Purpose Environment env:dev, env:test, env:stage, env:prod Used to identify the environment type Billing bu:bigbu cc:sales region:emea One ore more tags used to allocate costs Application app:bigapp svc: jenkins One or more tags to define the application or service Compliance dataresidency:germany compliance:pii compliance:hipaa One or more tags to define and compliance requirements Optimization schedule:24x7 schedule:12x5 maxruntime:14days One ore more tags to use in automated optimization
- 11. • Tag all types of resources that you can • Resource naming conventions not enough • Use the same tags for all clouds/environments • Be exact • Spelling, spaces, punctuation, upper/lowercase matters • Use existing automation tools to apply tags Tagging Tips
- 12. • Sample rollout process • Stage 1: Define and communicate required tags • Stage 2: Scheduled reports on missing tags by team/app • Stage 3: Instant alerts on missing tags with 12 hour shutdown warning • Stage 4: Instant alerts and shutdown Tagging Rollout Process
- 13. • Account-Based vs. Tag-Based • Each billing unit has own account(s) • Shared accounts with costs allocated via tagging • Purchase/Allocation of Discounts • Centralized purchase – everyone saves • Allocate savings proportionally (blended rates) • De-centralized purchase – buyer saves • Allocate savings to buyer (unblended rates) • Handling “upfront” payments for purchase commitments • Allocate at purchase time based on current usage levels • Amortize and allocate based on actual usage • Markups for IT overhead Cloud Cost Allocation Considerations
- 14. • Combine Push and Pull • Schedule automated reports by team/group • Enable ad-hoc access to latest and greatest cost data • Frequency • Typically weekly • Daily for variable or fast changing environments • Highlight anomalies • Significant changes outside normal ranges Reporting
- 15. • When to forecast • Annual budget cycle • Rolling forecast cycle • When budget is exceeded for x months • For new applications/projects/initiatives • Approaches to Forecasting • Projected growth patterns • Change cloud provider • Change instance types • Other “what-if” Forecasting
- 16. • Set Budgets • Use forecasts to create budgets • Budget Alerts • Alert when current or projected monthly spend exceeds budgets • Provision-time controls • Soft limits • alert if over budget • request approval before launch if over budget • Hard limits • prevent launch if over budget Budget Enforcement
- 17. • Tag requirements • Allowed clouds/regions/services/instance types • Allowed schedules • Automated cloud placement by cost/compliance/requirements • Budget enforcement Automated Provisioning Policies
- 19. RIGHTSIZING
- 20. The High Cost of Overprovisioning m3.xlarge $.266 $2330/year m3.large $.133 $1165/year m3.medium $.067 $587/year Save 50% Save 75%
- 21. 0 10 20 30 40 50 60 70 80 90 100 CPU Util% Mem Util% Don’t Provision to the Peak Don’t provision to the peak! Provision for normal loads and auto-scale for peak
- 22. Underutilization is Rampant Memory utilization CPUutilization 20-40%: Size -1 <20%: Size -2
- 23. Underutilization is Rampant Memory utilization CPUutilization High CPU utilization Low memory utilization Low CPU and memory utilization Custom VMs Downsize
- 24. Custom Sizing Example (GCP) 4 vCPU 20 GB You Need Standard VM Custom VM GCP Cost = $.280/hr GCP Cost = $.163 Savings = 42% 8 vCPU 30 GB 4 vCPU 20 GB
- 25. IDLE RESOURCES
- 26. Idle Instances
- 27. Unused Resources
- 28. Don’t Forget Storage Volumes Snapshots Attached Unattached
- 29. Finding Unattached Volumes Cost • AWS = $200+ per month • Azure = $40-120 per month • Google = $80-340 per month
- 30. DEV ENVIRONMENTS
- 31. Development Environment Usage Hours Mon Tue Wed Thu Fri Sat Sun 00:00 11:59 24x7 168 hours 12x5 60 hours 35%
- 32. Development Environment Shutdown Dates Needed for 3 days Left running for 4 days Left running for 7 days Left running for 14 days 25% waste 57% waste 79% waste
- 33. A Little Waste Adds Up Typical m3.large ($.133) 24x7 16 days Optimized m3.medium ($.067) 12x5 14 days $51.07 $11.26 78% less $102,140 $22,520 Per launch 100 devs * 20x/yr Save $79K
- 34. Scheduling Workloads in RightScale
- 35. • Training • Demos • Sandboxes • Test • Staging Other Temporary Workloads
- 37. Regional Differences in AWS Example Region Location Instance Size Hourly Cost Cheaper Region Location Hourly Cost % savings us-west-1 NorCal m3.large $0.15 us-west-2 Oregon $0.13 14% eu-central-1 Frankfurt m3.large $0.16 eu-west-1 Ireland $0.15 8% ap-southeast-1 Singapore m3.large $0.20 ap-southeast-2 Sydney $0.19 5% ap-northeast-1 Tokyo m4.large $0.17 ap-northeast-2 Seoul $0.17 5% Expensive Region Monthly Spend Cheaper Region % savings Monthly savings us-west-1 $5,000 us-west-2 14% $700 eu-central-1 $0 eu-west-1 8% $0 ap-southeast-1 $2,000 ap-southeast-2 5% $100 ap-northeast-1 $0 ap-northeast-2 5% $0 $800
- 38. Regional Differences in Azure Example Region Location Instance Size (Linux) Hourly Cost Cheaper Region Location Hourly Cost % savings East US Virginia D1v2 $0.07 East US 2 Virginia $0.06 12% North Central US Illinois D1v2 $0.07 South/West Central US Texas $0.06 12% Central US Iowa D1v2 $0.07 South/West Central US Texas $0.06 12% West US California D1v2 $0.07 West US 2 $0.06 12% Canada Central Toronto D1v2 $0.08 Canada East Quebec City $0.07 9% West Europe Netherlands D1v2 $0.08 North Europe Ireland $0.07 14% East Asia Hong Kong D1v2 $0.11 Southeast Asia Singapore $0.09 15% Japan East Tokyo D1v2 $0.11 Japan West Osaka $0.09 13% Australia East NSW D1v2 $0.09 Australia Southeast Victoria $0.08 7%
- 40. Superseded Instance Types Cloud Provider Previous Generation Current Generation Size mapping % savings (East) AWS T1 T2 same size 35% AWS M1 M3 same size 24% AWS M2 R3 downsize one 32% AWS C1 C3 same size 60% AWS CR1 R3 same size 24% AWS HS1 D2 downsize one 40% Azure D* D*v2 same size 4%
- 42. Your Reality is Constant Change Family A 2 vCPU 4 GB Family C 8 vCPU 8 GB Now Family A 2 vCPU 4 GB Family B 2 vCPU 8 GB Family B 2 vCPU 8 GB Family A 2 vCPU 4 GB Family A 4 vCPU 8 GB Future Family A 2 vCPU 4 GB Family B 2 vCPU 8 GB Family B 1 vCPU 4 GB Family B 1 vCPU 4 GB Family A 2 vCPU 4 GB Family A 2 vCPU 4 GB
- 43. AWS RIs Azure CPP Google SUD/ CUD IBM Monthly Length of commitment 1 or 3 years CPP Compute pre-purchase for 1 year (must have EA) SUD: No commitment Monthly: Commit by month Payment No Upfront Partial Upfront All Upfront All Upfront No Upfront By month Range of discount levels RI (1Y) 24-58% RI (3Y) 32-75% CPP 19-63% SUD: Up to 30% CUD: 37% (1Yr) or 55% (3Yr) Monthly: About 10% Commitment Discount “Grouping” *Region + Instance type + OS + **Network *Regional benefit **Network can be modified Datacenter + Instance type + Instance size + OS Has time flexibility Region + # vCPUs + # GBs RAM *Across any instance size/type Any specific individual resource Comparing Commitment Discounts by Cloud
- 44. Commitment Discounts Differ in Flexibility m3.lg m3.lg m3.lg m3.lg m3.xl m3.xl m3.2xl c4.lg c4.xl c4.2xl c4.2xl r3.xl r3.4xl r3.xl D2 v2 D2 v2 D3 v2 F3 v2 G4 v2 F3 v2 D2 v2 D2 v2 G3 v2 n1-std-2 n1-std-2 n1-std-2 n1-std-2 n1-std-4 n1-std-4 n1-std-8 n1- highmem -4 n1- highmem -8 n1- highmem -8 n1- highcpu- 4 n1- highcpu- 16 n1- highcpu- 4 Region Datacenter Region AWS RIs Azure CPP Google CUD
- 45. Azure Gives Time Flexibility D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 D2 v2 Day 1 Day 2 Day 3 Day 1 Day 2 Day 3 Day 1 Day 2 Day 3 Buy blocks of 744 hrs/month of an instance type/size/OS/region and use any time in the month
- 46. About Google Sustained Use Discount (SUD) • No commitment. The more you use an instance family during the month, the higher the discount. Usage Level % of Billing Cycle Incremental Rate % of On-Demand Baseline Sample Rate n1-standard-1 Total Cost 0-25% 100% $0.050 $9.00 25-50% 80% $0.040 $7.20 50-75% 60% $0.030 $5.40 75-100% 40% $0.020 $3.60 Monthly Cost at 100% usage 30% discount $25.20
- 47. • Commit to # of vCPUs and GBs of RAM • 1 yr or 3 yr • Can be used for any instance type or size in a region • SUD still applies for non-committed use Google CUD is Based on Family/Size
- 48. Think “Commitment Discount” Coverage 100 instances 50 Instances under CD 50% CD coverage 50% On-Demand pricing
- 49. Usage/Cost Pattern for a Commitment Group 48 Production and 24x7 dev usage Weekday dev usage
- 50. Think “Commitment Discount” Coverage 49 Target CD Coverage may range from 50-90% Depends on level of change planned and flexibility of commitments
- 51. You Need to Monitor “CD” Utilization
- 52. • Interested in managing cloud costs and optimizing spend? • [email protected] Q&A