Case Study: HCL Technologies On Capacity Planning for Cloud and Virtualized Environments
1 like2,922 views
The document discusses the capacity management process and its evolution from traditional models to modern cloud solutions. It highlights the need for balanced capacity provisioning that anticipates demand fluctuations to improve efficiency and performance while minimizing costs. Key factors include virtualization, ongoing capacity management, and proactive resource optimization to meet business dynamics.
Case Study: HCL Technologies On Capacity Planning for Cloud and Virtualized Environments
- 1. ca Opscenter Case Study: HCL Technologies On Capacity Planning for Cloud and Virtualized Environments Navin Sabharwa HCL Technologies Practice Head Public Cloud, Cloud Mgmt, Automation, Analytics, DevOps OCT60S #CAWorld
- 2. 3 Objective o The goal of the Capacity Management process is to ensure that cost-justifiable IT capacity in all areas of IT always exists and is matched to the current and future agreed needs of the business, in a timely manner.
- 3. 4 Traditional Capacity Model Concerns A pessimistic approach as there was focus on providing highest possible unit of capacity to support applications to run desirably in peak hours. In off peak hours, procured resources sat idle and were underutilized. On the other hand, constrained resources would lead to overutilization of available resources leading to performance issues. There was a lack of balance between demand and capacity because capacity requirements did not flow from business level to service level and then to component level. This was a short term approach (incident based) with focus on component capacity. A lack of proper planning resulted from an absence of an inter-process relationship for capacity planning and forecasting.
- 4. 5 New Age/Cloud Capacity Solution The focus is on providing the smallest possible unit of capacity to support an application. The smallest possible unit for capacity has reduced from a complete Hardware Stack to a Flexible Virtual Server which can be provisioned and de-commissioned based on need. Virtualization in cloud computing allows for workload migration and optimum capacity utilization. Cloud computing provides scalable infrastructure which can be provisioned in minutes as compared to weeks in traditional environments. Ensure services meet their performance targets with cost economies and flexibility to the consumer to scale capacity.
- 5. 6 Delivery Process Interdependencies
- 6. 7 Support Process Interdependencies
- 7. 8 Iterative Capacity Management Ongoing/Iterative capacity management procedures are required by service providers when existing service needs to be implemented and optimized for business and performance.
- 8. 9 Capacity Management Meeting Demand Capacity must be able to intelligently tune itself according to criticalities that may arise due to business dynamics, seasonal and irregular variations.
- 10. Capacity Demand Coupling 11 The ability of the cloud provider to anticipate the rise and fall in demand is the key to being a successful cloud provider. In case of under capacity, the cloud consumers will not be able to provision resources or the performance SLAs will suffer; resulting in customer dissatisfaction and financial loss.
- 11. Resource Reclamation Process 12 This process identifies workload inefficiencies and validates the same. Identified underutilized resources are reclaimed. Environment is monitored for further resource and cost optimization.
- 12. Modeling and Forecasting 13
- 13. Capacity Forecasting 14 In optimistic capacity forecasting, we would reserve more capacity for redundancy and other factors.
- 15. Workloads 16
- 16. Hourly CPU Load Optimized by time of day 17 14 12 10 8 6 4 2 0 25% Savings 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Load Hour
- 17. Monthly Optimization Optimized during a month 18 RDS DB Servers 75% Savings Daily CPU Load 1 3 5 7 9 11 13 15 17 19 21 23 Days of the Month
- 18. Worst Case Scenario – AWS CloudFront http://www.reviewmylife.co.uk/blog/2011/05/19/amazon-cloudfront-and-s3- 19 maximum-cost/ Author calculated maximum possible charge: – Used default limit of 1000 requests per second and 1000 megabits per second – At the end of 30 days a maximum of 324TB of data could have been downloaded (theoretically) – $42,000 per month for a single edge location – CloudFront has 30 edge locations
- 19. Stories And Lessons Learned Anecdotal user experience 20 – Personal website hacked by file sharers – Received bill for $10,000 Note: AWS only charges for data out – All data transfer in is at $0.000 per GB – Mitigates costs – if you don’t respond to requests, it doesn’t cost you anything
- 20. CA Capacity Management Increase Efficiency, Assure Delivery and Reduce Costs With Confidence 21 DECISION SUPPORT FOR IT INVESTMENTS Model growth Assess capacity efficiency across IT Identify utilization impact to business services PREDICTIVE ANALYTICS Anticipate potential issues before they impact the customer experience Strategic Business Value
- 21. What is CCM in real-life? 22
- 22. Looking Closer…Increased Visibility 23 Capacity across IT Capacity across facilities Unified view
- 23. Looking Closer…Increased Visibility 24 Capacity across IT Capacity across facilities Unified view Monitor breaker used by VM clusters View IT Utilization on same clusters Alarm = power exceeds capacity rating Power issues visible before they present a problem
- 24. Looking Closer…Increased Visibility 25 Capacity across IT Capacity across facilities Unified view Indication of productive and unproductive power usage Unproductive power = potential to reduce IT Equipment Reduced IT equipment = increased efficiency & less power
- 25. Looking Closer…Improved Provisioning Capabilities 26 Capacity analysis - What ‘What-if’ analysis - Where Deployment - How
- 26. Looking Closer…Reduced CapEx and OpEx 27 Optimization of software Optimization of hardware Risk mitigation
- 27. For More Information 28 Insert appropriate screenshot and text overlay from following “More Info Graphics” slide here; DevOps ensure it links to correct page To learn more about DevOps, please visit: http://bit.ly/1wbjjqX
- 28. For Informational Purposes Only © 2014 CA. All rights reserved. All trademarks referenced herein belong to their respective companies. This presentation provided at CA World 2014 is intended for information purposes only and does not form any type of warranty. Content provided in this presentation has not been reviewed for accuracy and is based on information provided by CA Partners and Customers. Terms of this Presentation 29