![Assure Application QoS: Right-sizing a Virtual Machine
VM Prices VM ROI >> 0 VM Prices
[ROI = Rev. – Exp.]
Application VM VM
Workload in Resource Resource
a VM Utilization Utilization
VM VM VM VM VM
Hypervisor](https://image.slidesharecdn.com/vmturboslidedeckfinal-13439975318787-phpapp02-120803074159-phpapp02/85/Vm-Turbo-Slide-Deck-18-320.jpg)
Vm Turbo Slide Deck
- 1. VMTurbo Overview Intelligent Workload Management for Cloud and Virtualized Infrastructures
- 2. VMTurbo Overview • Founded in 2009 • VMTurbo Operations Manager launched in September 2010 • An award-winning Intelligent Workload Management solution • 8,000+ users; 160+ customers worldwide • Growing Channel Ecosystem Our Focus: Rethink the Approach to IT Operations Management
- 3. Virtualization Introduces Broad Set of Management Decisions 1 VM and Application Sizing: Determining the number of instances, resource allocations, and appropriate VM VM VM VM VM VM VM prioritization to meet business demands 2 Initial and Ongoing Placement Decisions: Determining where workloads should run – or be relocated to – across the available compute, storage and network resources 3 Capacity Management: Determining how much capacity to configure today, as the environment grows, and during hardware refresh cycles To be effective, these decisions must be made holistically
- 4. And the Number of Control Points Increases Exponentially • Multiple layers • Multiple entities in each layer • Dozens of controls for each entity VM VM Hypervisor • Thousands of control points • Interdependencies galore • An overflow of metrics across the board 4
- 5. Rethinking the Approach to IT Operations Management Iterative Advancements in the “State of the Art” Collect Threshold Alert Suppress & Correlate Learn Predict Stop Collecting – Start Controlling Stop Alerting – Start Preventing VMTurbo Confidential - Do Not Distribute 5
- 6. A Smarter Way: Intelligent Workload Management • Changes the way virtualized data centers are controlled • Provides a closed loop system that automates the decision-making process • Focuses on 2 key business requirements Efficiently Assuring utilizing the Application underlying Performance infrastructure
- 7. Intelligent Workload Management is About Maintaining the Optimal Operating Zone ms Doing this in a virtual Optimal Operating environment with Delay Zone • Workload fluctuations • Shared resources • Interdependencies QoS Guarantee • Configuration changes Efficiently • New applications, Assuring utilizing the users and hardware Application virtualized 0% Utilization 100% Performance infrastructure …Requires constant tuning & optimization
- 8. How Is VMTurbo Different? Economic Scheduling Engine Market-based Approach: • Abstraction representing the IT infrastructure as a marketplace • Applications & physical resources depicted as buyers & sellers Combined with Analytics: • Driving intelligent decision-making for resource allocation • Rapidly adjusting to changes in workload demand & capacity Result: Executable Actions Which Prevent Problems and Drive a Healthy & Efficient Infrastructure
- 9. What VMTurbo Operations Manager Does Discover Converge Control Simulate • Virtual appliance • Immediately delivers • Continuously profiles and • Models the effect of installs in minutes actionable tunes environment changes to the • Single instance recommendations • Automates resource environment manages vSphere, • Drives the environment to allocation and placement • Projects future demand XenServer, Hyper-V healthy state decisions based on trends • No “self-learning” & • Assures QoS and • Addresses contention • Delivers plans to no new databases maximizes efficiency before problems occur accommodate growth VMTurbo Confidential - Do Not Distribute 9
- 10. Intelligent Workload Management Delivers IT Automation & Control VMTurbo Operations Manager Key Benefits VM Placement Configuration Constraints VM Configurations • Service & application assurance VM Relocation Real Time Operations • Operational & labor Application Resourcing Economic efficiency Scheduling … Engine • Improved resource Compute Resource Performance Server Packing utilization Metrics Capacity Storage Resourcing • Simplified virtualization Service Level Application Sizing management Intelligent Priorities Planning Server Refresh Planning IT Automation & Control Abstraction, Analysis & Automated Actions
- 11. A Smarter Way to Control Your Virtualized Data Center Close the Loop • Stop collecting & alerting – Start preventing & controlling Improve Efficiency • Increase utilization through iterative planning & capacity management Simplify & unify virtualization management Take the Challenge • Minutes to value – Free – www.vmturbo.com/download 11
- 13. Comparing Approaches to Operations Management Attribute Competitive Approach VMTurbo Approach Inbound alert mechanisms require administrators to Closed-loop control system that continuously Automation & interpret data, determine fix, and then enact identifies areas of congestion and automatically Control remediation applies the proper action to remediate Prevents problems by continuously tuning the Threshold-driven, alert-triggered approach to problem Problem Handling environment to maintain health and avoid avoidance bottlenecks Coarse-grained, aggregate metrics taken in as small as 5- Real-time metrics drive pricing algorithm for Granularity minute chunks immediate response to escalating contention Based on available “first fit” decision by VM Holistically assess the entire environment to Placement configuration and do not take other factors into determine the most optimized configuration Decisions consideration inclusive of the incoming workload Planning Capacity-driven analysis based on workload Performance-driven analysis based on workload Methodology configuration characteristics and holistic view of entire system
- 14. Comparing Approaches to Operations Management (2) Attribute Competitive Approach VMTurbo Approach Unified virtualization management, including Closed system for hypervisor vendors; Platform Support vSphere, XenServer, Hyper-V, CloudStack, vCloud 3rd-party vendors require separate appliances Director – all from a single appliance Purpose-built system designed to control the virtual Assembled technologies with multiple databases and Architecture environment with market analytics in an automated user interface overlays fashion Provides actionable, intelligent control within Requires time to collect metrics and “learn” behaviors Time-to-value minutes of deployment – converges environment to before delivering data for analysis healthy state, and maintains it there Per-VM license packs create density-tax and make Per-socket licensing provides lower ongoing costs Licensing Model budgeting complicated, costs range over $2k/host and more certainty in budgeting, costs ~$1k/host Drill-down analysis requires high-cost, experienced Automated decisions and preventive measures TCO admins; resource intensive troubleshooting and SLA eliminate “fire drills,” reduce operator overhead, and exposure not resolved assure application performance
- 15. Platform Approach to Intelligent Control Across the Data Center Prevent Detect Resolve Control Plan Optimize Alert Report Automation & Presentation Layer Analytic Layer – Economic Scheduling Engine Abstraction Layer – The Market Mediation Layer vSphere Compute Storage App Delivery Controllers vCloud XenServer Director CloudStack Hyper-V 15
- 16. Economic Scheduling Engine – A Market Approach Commodities Market – Trading Compute Abstraction Resources VM VM Hypervisor VM Buyers Sellers Buyers Sellers Buyers Sellers Buyers Sellers Buyers Sellers 16
- 17. Economic Scheduling Engine – Analytics Driven by Pricing ms Delay Price Abstracts QoS Guarantee 0% Utilization 100% 0% Utilization 100%
- 18. Assure Application QoS: Right-sizing a Virtual Machine VM Prices VM ROI >> 0 VM Prices [ROI = Rev. – Exp.] Application VM VM Workload in Resource Resource a VM Utilization Utilization VM VM VM VM VM Hypervisor
- 19. Business Challenges Resolved with This Approach • Preventing performance issues related to workload delays or outages • Remediating building contention – maintaining/improving “health” • Reducing the risk of production outages • Improve utilization and intelligently plan for VM growth • Increase VM density – reduce cost • Plan accurately for new projects, organic growth and hardware refresh • Reduce CAPEX on servers & associated software • Improve the efficiency of operational resources • Drive automation and control across the environment – better admin/server ratios • Reduce trouble-tickets and OPEX costs associated with the environment 19
- 20. Product Demonstration VMTurbo Confidential - Do Not Distribute 20
Editor's Notes
- #3: Product has been in market for just over 18months stemming from innovation from the founding team at SMARTS – acquired by EMC in 2004. Deep rooted history and IP in building management technologies for large, complex, distributed IT environments.Since our product launch in in Oct 2010, the first 6 months focused on proving out the technology – but we now have a mature and award winning IWM solution. We have seen huge adoption growth in past calendar year adding more than 100 customers since summer of last year.We have more than 4,000 users of our free monitoring, reporting and alerting offering and over 150 paying customers using our intelligent, automated data center control product – VMTurbo Operations ManagerCustomers range from the worlds biggest cloud providers, to large enterprises and enlightened IT departments at mid-sized businessesAll are unified by the same desire to Think Differently about their approach to IT Operations Management