Bringing the Private Cloud to the Data Center Matt Eastwood Group Vice President Enterprise Platforms Research September 2010 Copyright 2010 IDC. IT in the Cloud is a Journey… The Market Has Been Cloud Washed… What’s Next? Enjoy the Ride Copyright 2010 IDC. Cloud Computing Defined Cloud Computing – consumer and business products, services and solutions delivered in real-time over the Internet including: – Standard shared service (built for a market, not a single customer) – Solution-packaged (“turnkey” integrating appropriate resources) – Self-service – Elastic scaling (dynamic and fine-grained) – Service Metering (use-based pricing) – Ubiquitous Network Access (accessible via the Internet) – Browser Based (standard UI technologies) – Published API (common Internet APIs) Copyright 2010 IDC. Deployment Models Public: Open to a largely unrestricted universe of potential users; designed for a market, not a single enterprise Private: Designed for, and access restricted to, a single enterprise (or extended enterprise); an internal shared resource, not a commercial offering; IT Org is the “vendor” of the shared/std service to its users Cloud Adoption: Evolutionary vs. Revolutionary Private Cloud Datacenter ATraditional Look Back to 2005 Private IT Dynamic IT Dynamic Cloud Migration Existing ‘Custom’ Copyright 2010 IDC. Public Cloud PublicxSPs xSPs Cloud Elastic Scaling New New ‘Bursting’ Agenda Market Trends Attitudes Towards 'The Cloud' Cloud and Convergence Essential Guidance Copyright 2010 IDC. The “New Normal” Shifting Business Goals Impact Technology Goals Q. Prioritize the following business goals as they relate to your organization by allocating 100 points among them. 100% NC +4 90% 80% +3 70% 60% +6 50% Speed time to market Increase market share Improve quality/ accuracy Increase revenue 40% +3 30% Increase customer sat. 20% -16 10% Reduce costs 0% Jan. '07 Copyright 2010 IDC. Oct. '08 Sept. Jan. '09 '10 Source: IDC Enterprise Platforms Surveys A more balanced view of end-user business goals is returning to market With virtualization, focus has been on cutting costs Customers are again beginning to look for additional business benefits from IT Technology will clearly play an important role in future business initiatives Worldwide Server Market, 1996-2010 +8% -30% +23% -27% +$5.5B -$20B +$11B -$17B Revenue (%) 100% 14% Other Servers 90% ‘97-’08 Change 45% 80% 31% 70% Unix Servers 60% 50% Other -31 Pts Unix 32% 40% -1 Pts 30% 20% x86 Servers 23% 10% 55% +22 Pts 0% '97 '98 '99 '00 '01 '02 '03 Source: IDC Quarterly Server Tracker, Q4 2009 Copyright 2010 IDC. '04 '05 '06 x86 '07 '08 '09 '10 WW x86 Server Shipment Forecast: x86 Servers Continue to Accelerating YOY Shipment Growth 4Q Rolling Shipments 30% 9,000,000 Q108 Q208 Q308 Q408 Q109 Q209 Q309 Q409 Q110 Q210 Q310 Q410 8,000,000 20% 7,000,000 10% 6,000,000 0% 5,000,000 -10% 15% 100% -20% -30% Q308 Q309 -1.7M -21% Q309 Q410 +1M Units +16% 45% 4,000,000 3,000,000 2,000,000 Q409 YOY Growth 85% 4Q Rolling Shipments -40% 1,000,000 0 Server Deployments are Increasing Rapidly on Refresh Copyright 2010 IDC. Source: IDC Quarterly Server Forecast. June 2010 The Technology Catalyst: WW Growth 2009-2013 15 10 Efficiency Complexity Elastic Scaling Off Premise 12.4x 8.4x 4.7x 5 1.3x 2.3x 2.5x 2.7x 0 Servers (M) Copyright 2010 IDC. Blades (M) Cores (M) VMs (M) Data Transactions (TBs) 10G Ports Server Installed Base: The Evolution From Physical to Virtual ’09-’13 CAGR Servers 100,000,000 75,000,000 50,000,000 WW Virtual Server IB 31.5% 25,000,000 WW Server IB 0.6% WW Server Shipments 6.5% 0 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 '12 '13 Virtualization Leaves its Mark and Primes the Market for Change Copyright 2010Source: IDC. IDC Server Virtualization 2009 Virtualization is the Foundational Platform for the Datacenter • More than half of all workloads (51%) will be virtualized by the end of 2010 • Two-thirds (69%) by 2013 • Only 12.8% of all physical servers are virtualized in 2009 WW Installed Workloads Virtualized 75% 50% 25% • VM densities continue to rise predictably • Averaging 6 VM’s per physical server in 2009 and 8.5 in 2013 Copyright 2010 IDC. 0% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2.1% 12.8% 22.3% Servers Servers Servers WW Server Virtualization Shipment Forecast, 2005-2013 1.4M VMs 15M VMs VM Cross Over 8.2 7.7 7.1 3.0 2005 3.3 4.2 2006 2007 5% Shipments Copyright 2010 IDC. 5.3 2008 6.1 2009 2010 2011 2012 2013 21.6% Shipments New Economic Model for the Datacenter Shifts to Automation Tools are a Requirement WW Spending on Servers, Power and Cooling, and Management/Administration 56 million virtual machines by 2012 Virtualization Management Gap Copyright 2010 IDC. Virtualization is Foundation for Cloud Consolidate Virtualize Automate Provision Cloud 10% 20% Self Provision 90% Metering & Chargeback Metering & Chargeback Mobility Mobility Mobility Hypervisor Hypervisor Hypervisor Hypervisor 45% Sites 18% Market 90% of Virt. Sites 20% of Virt. Sites 10% of Virt. Sites 18% Copyright 2010 IDC. Cloud Internal vs. External Source: IDC Server Virtualization 2009 IT Optimization by the Numbers: Houston We Have a Problem Q. What were the top three areas of savings? (N=400) % of Responses % of Annual Spend 50% 40% -15% -5% -9% -8% +20% +17% 30% 20% 10% 0% RE Copyright 2010 IDC. Server System Staff HW SW Costs Drop 25% but 80/20 Rule Still Applies Support P&C Source: IDC Server Virtualization 2009 Who’s Embracing the Cloud Q. Rate utilization of public cloud over next 3 years (1-10 Scale)? Total Cloud is divisive concept with SMB and mature virtualization site favoring <500 Employees 500-4,999 5,000+ SMB’s consuming SaaS applications Startups and enterprises with large scale Web infrastructures <15% 15-<50% >50% Virtualized Enterprise developers with one-off projects 3.00 3.25 3.50 3.75 4.00 4.25 4.50 4.75 Copyright 2010 IDC. Compute intensive tasks Source: IDC Enterprise Platforms Survey, Jan 2010 (N=255) Workloads in the Cloud Q. Which applications will you run in Cloud in 3 years: Bus. Processing 18% DSS 16% VOIP 13% IT Infra 14% App. Dev. 8% Tech. 31% Collaboration 41% email 21% Web Infra 0% Copyright 2010 IDC. 25% Agnostic Private 32% 50% Workloads remain the critical pivot point in the datacenter Virtualization is hot and it impacts cloud adoption profiles Users with high rates of business change favor cloud models 75% Source: IDC Enterprise Platforms Survey, Jan 2010 (N=255) Server Workloads Topology 18 Workloads in 7 Workload Categories Business Processing Business Processing Web Infrastructure Infrastructure ERP Streaming Media CRM Web Serving OLTP Batch Collaborative Collaborative E-Mail Workgroup Decision Support Decision Support Data Warehousing/Mart Data Analysis/Mining Copyright 2010 IDC. App. Development Development App. Infrastructure IT Infrastructure File & Print Networking Proxy/Caching Security Systems Management TechnicalR&D Industrial Source: IDC Server Workloads 2009 Three layers of cloud computing Workloads email ERP CRM OLTP Collaborative IT Cloud Services Cloud Applications Software as a Service (SaaS) Decision Support App. Dev. Caching Web Platform as a Service (PaaS) Streaming Security Networking Technical Cloud Platform File Sys Mgmt Cloud Infrastructure Infrastructure Services (IaaS) Server Copyright 2010 IDC. Storage Network Platforms Actual Mileage May Vary: Migration Starting Point Data Analysis $16 Billion $44.9 email Data Mining Streaming Web App. Dev. Collaborative 39% 41% $14.0B $17.6B PaaS SaaS ERP CRM OLTP 22% Technical Caching Security Networking Copyright 2010 IDC. $13.3B IaaS Sys Mgmt File Serving Source: IDC Server Virtualization 2009 Cloud Implementation Help Q. Who will help build out your private cloud? Larger companies and leaders favor internal IT Internal IT Department Virtualization Vendor Server Vendor Smaller orgs and laggards favor server vendors Networking Vendor Systems Integrator Storage Vendor Virtualization vendors are broadly favored ISV Systems Management Outsourcer Local VAR 0% Copyright 2010 IDC. 20% 40% 60% 80% Those planning private clouds are most likely to favor Internal IT Source: IDC Enterprise Platforms Survey, Jan 2010 (N=255) Cloud Leadership Comparing mature and understood (virt) with unknown (cloud) The evolution to private cloud will be led by traditional IT Traditional IT suppliers lag in Public cloud Larger orgs name IBM SMB favors Google, Amazon and Salesforce Copyright 2010 IDC. Q. Who is the leading vendor? 100% Don't Know Other Salesforce.com 75% Amazon 50% Google 50% Cisco 45% 25% 55% Microsoft HP IBM VMWare 0% Virt Private Public Source: IDC Enterprise Platforms Survey, Jan 2010 (N=255) Attitudes Towards Convergence Q. Most appeal w/ convergence? 3.6 3.3 3.9 100% Don't Know Other • The sever will be the focal point for IT convergence 75% FCoE • Larger companies, leaders and leading virtualization adopters favor most Storage & Networking 50% Server & Network 25% Server & Storage Virtualization & Automation 0% Total (N=255) Copyright 2010 IDC. Laggards (N=131) Leaders (N=124) • End-users rate likelihood of converged adoption at 3.6 of a 10 scale • Service providers more likely to favor network led convergence Source: IDC Enterprise Platforms Survey, Jan 2010 (N=255) New Business Cycle for IT Business Value Convergence Reigns for the next 10 years Transactional applications and Database ERP, Analytics and Datamarts Converged Infrastructure & Private Clouds Web Collaborative Application Development Public Clouds 1985 Copyright 2010 IDC. 1990 1995 2000 2005 2010 2015 File/Print and Networking Intersection Technology Offerings Traditional IT Outsourcing Hosting Public Cloud Private Cloud Traditional Servers Converged Infrastructure Homogeneous HeterogeneousInfrastructure: Infrastructure: multi-node computing designresource including Server, storage and network processor, memory, network, pools assigned as needed to storage, business P&C in a single chassis services Copyright 2010 IDC. Micro Servers Market Segments Continue to Shift: Understanding the Impact of Cloud Units % Server Spend 100% 75% 50% PrivateCloud Cloud Private – Bladed and Virtualized – Consolidation Play 81% 70% – Existing Workloads 93% PublicCloud Cloud Public – Homogeneous 17% 20% – Dense/Low Cost 2% 10% – New Workloads 25% 0% 6% 1% 2009 2010 Copyright 2010 IDC. TraditionalIT IT Traditional – Heterogeneous – Legacy & Proprietary 2011 2012 2013 2014 Cloud and Convergence, 2013 The Whole is Smaller than the Parts 2013 Mkt CAGR (10-13 ) Not $144B But $123B Not 3.2% But -2.0% Security Networking Equipment Storage Systems $62 Billion $32 Billion Streaming eMail Web $11B Caching Server Hardware File&Print Networking $50 Billion Workgroup Copyright 2010 IDC. $10B Cloud and Convergence, 2013 A View of a Bifurcating Server Market $52B: Bifurcation Small Site 19% Scale-out 9% Datacenter 34% Copyright 2010 IDC. Non x86 38% An Alternative View (or is it) Workload Optimization $52B: Optimization OLTP/DB 15% BA 12% BI 12% Infrastructure 61% Copyright 2010 IDC. Essential Guidance CIOs pay attention to time, money and people Downturns drive inflections and change on recovery Workloads are the critical pivot point for cloud and convergence decisions Many political barriers remain in the DC User demographics change constantly Copyright 2010 IDC. Questions Matt Eastwood meastwood@idc.com Twitter: matteastwood +1 (508) 935-4503 Copyright 2010 IDC.