Google’s Committed Use Discounts vs. AWS Reserved Instances and More Ways to Save with Google
2 likes645 views
The document discusses cost optimization strategies for cloud services, focusing on Google's committed use discounts compared to AWS reserved instances. It highlights potential savings of 30-45% through the use of preemptible VMs and custom machine types, alongside real-world usage examples. Various use cases illustrate the benefits of proper rightsizing and leveraging cloud pricing models to avoid wasted expenditures.
Google’s Committed Use Discounts vs. AWS Reserved Instances and More Ways to Save with Google
- 1. GOOGLE COMMITTED USE DISCOUNTS VS. AWS RESERVED INSTANCES AND MORE WAYS TO SAVE WITH GOOGLE
- 2. • Manish Dalwadi • Product Manager, Google •Michael Crandell • CEO, RightScale Panelists
- 3. • Google’s approach to cloud pricing • Ways to save money on Google • New Committed Use Discounts • Real-World Examples • AWS Reserved Instances vs. Google Committed Use • Preemptible Virtual Machines • Custom VMs •Q&A Agenda
- 4. Cost Optimization is Top Initiative in 2017
- 5. RightScale Helps you Save Money Typical Savings 30-45%
- 7. Self-Estimated Wasted Spend Additional Wasted Spend Typically Found by RightScale 30-45% of Cloud Spend Is Wasted Source: RightScale 2017 State of the Cloud Report
- 8. The Three Biggest Areas for Waste Not Leveraging Discounts Running Dev Instances 24x7 Oversized Instances
- 9. How Google Helps Committed Use and Sustained Use Discounts Preemptible VMs Custom Machine Types
- 10. USE CASE 1: OPTIMIZING YOUR DISCOUNTS
- 11. • We know we can save with RIs, BUT: • We have underutilized instances that I need to downsize • We are moving to Docker and that will change what I need • We have dev instances that are changing all the time • We need to re-architect that system • We may need to change our instance families • There may be new instance types coming • …and more The “Yeah, Buts” of Reserved Instances
- 12. Your Reality Is Constant Change 2 vCPU 4 GB 8 vCPU 8 GB Now 2 vCPU 4 GB 2 vCPU 8 GB 2 vCPU 8 GB 2 vCPU 4 GB 4 vCPU 8 GB Future 2 vCPU 4 GB 2 vCPU 8 GB 1 vCPU 4 GB 1 vCPU 4 GB 2 vCPU 4 GB 2 vCPU 4 GB
- 13. Example Usage/Cost Pattern Production and 24x7 dev usage Weekday dev usage * Data from real (anonymous) customer use cases
- 14. 90%** coverage of steady-state for Convertible RIs Example: Reserved Instances AWS *65% coverage provides for flexibility to change VM families and sizes used over time. 65%* coverage of steady-state for Standard RIs **Ability to change VM families and sizes enables higher coverage.
- 15. Example: Google Pricing *Ability to automatically change VM families and sizes enables higher coverage. *90% Committed Use Discount coverage of steady-state Sustained Use Discount coverage
- 16. AWS Customer Example with RIs Applied Rate Monthly Hours Monthly Spend % Savings RI (1 Yr Standard, No upfront) 65% coverage 650,000 $41,800 31% On-Demand 350,000 $37,800 0% Total 1,000,000 $85,900 21% RI (3 year Standard, Partial upfront) 65% coverage 650,000 $27,950 60% On-Demand 350,000 $37,800 0% Total 1,000,000 $65,750 39% RI (3 year Convertible, No upfront) 90% coverage 900,000 $60,300 38% On-Demand 100,000 $10,800 0% Total 1,000,000 $71,100 34% *Assumes all m4.large
- 17. Applied Rate Monthly Hours Monthly Spend % Savings Committed Use (1 year, no upfront) 90% coverage 900,000 $53,865 37% Sustained Use (100% usage level) 50,000 $3,325 30% Sustained Use (25-50% usage level) 25,000 $1,900 20% On-Demand 25,000 $2,375 0% Total Spend 1,000,000 $61,465 35% Committed Use (3 year, no upfront) 90% coverage 900,000 $38,475 55% Sustained Use (100% usage level) 50,000 $3,325 30% Sustained Use (25-50% usage level) 25,000 $1,900 20% On-Demand 25,000 $2,375 0% Total Spend 1,000,000 $46,075 52% GCP Customer Example with Discounts *Assumes all n1-standard-2 = $.095/hr 28% less than AWS 1 yr Standard RIs 35% less than AWS 3 yr Convertible RIs
- 18. Summary of AWS RI vs. Google Pricing
- 19. USE CASE 2: DEV/TEST WORKLOADS
- 20. Development Environment Usage Hours Mon Tu e Wed Thu Fri Sat Sun 00:00 11:59 24x7 168 hours 12x5 60 hours 35%
- 21. Sustained only Committed + Sustained 50/50 Committed + PVM 100% PVM Monthly Spend $67,450 $61,465 $40,733 $20,000 Additional savings vs. “Sustained only” 9% 40% 70% PVM Example *1M hrs/month; *95% of hrs at 100% usage, *2.5% at 50% usage *2.5% at On-Demand *Assumes all n1-standard-2 = $.095/hr
- 22. USE CASE 3: RIGHTSIZING WITH CUSTOM MACHINE TYPES
- 23. Underutilization Is Rampant Memory utilization CPUutilization
- 24. Underutilization Is Rampant High CPU utilization Low memory utilization Low CPU and memory utilization Custom VMs Downsize CPUutilization Memory utilization
- 25. Rightsizing 4 vCPU 20 GB You Need Standard VM Custom VM GCE Cost = $.266/hr AWS Cost = $.431/hr GCE Cost = $.155/hr Savings vs GCE std VM = 42% Savings vs AWS std VM = 64% 8 vCPU 30 GB 4 vCPU 20 GB *GCE Cost includes Sustained Use Discounts
- 26. Optimizing Google Cloud for ROI Committed Use and Sustained Use Discounts Preemptible VMs Custom Machine Types
- 27. See the blog: https://www.rightscale.com/blog/cloud-cost- analysis/aws-reserved-instances-vs-google-committed-use- discounts Contact: [email protected] Learn about Google: cloud.google.com/compute cloud.google.com/pricing/innovation Q&A