Next Steps for Federal Data Center Consolidation Initiative Presented to: GDI Conference April 28 2011 Creating Value, Delivering Results 2011 eglobaltech Incorporated. Tech, Inc. All rights reserved. 1
Agenda Speaker Introductions: Nick Mistry Senior Director, Cloud Computing Practice Jeff Vandervort Program Manager, GSA FDCCI PMO Presentation Agenda eglobaltech Corporate Background Federal Data Center Consolidation Initiative (FDCCI) Drivers for Data Center Consolidation Key Challenges for the Government Data Center Consolidation Methodology and Tools Consolidation Tools 2
eglobaltech Background Combining strong Federal Government experience with deep expertise in leveraging the value of technology and innovation Award-winning Federal Government EA programs at the Departments of Labor, Education and Energy Leading key Government-wide transformation initiatives including: o Federal Cloud Computing Initiative o Federal Data Center Consolidation Initiative o Federal Risk & Authorization Management Program (FedRAMP) Led the Federal Healthcare Information Sharing Pilot Lead authors of the Practical Guide for Federal Service Oriented Architecture and the Planning Guide and Roadmap for IPv6 Adoption Leading the Federal IPv6 Adoption Initiative 3
eglobaltech is Leading Government IT Transformation # OMB Action Item eglobaltech Leadership 1. Complete detailed implementation plans to consolidate 800 data centers by 2015 2. Create a government-wide marketplace for data center availability Supporting the Federal Data Center Consolidation Initiative (FDCCI) Developing TCO model Assisting in the development of govgrid a Government data center marketplace 3. Shift to a Cloud First policy Cloud Solutions Assessment / Planning Guides Gov-wide Cloud Services (IaaS, Email, PaaS) 4. Stand-up contract vehicles for secure Infrastructure as a Service (IaaS) solutions 5. Stand-up contract vehicles for commodity IT services Technical Requirements, Evaluation Criteria Development for IaaS BPA, awarded in Nov 2010 SaaS Email Procurement Technical Requirements Cloud Services Development for IaaS, PaaS and SaaS Apps.gov Development Support 6. Develop a strategy for shared services GeoCloud, Shared Services Pilot Implementation Architected shared service enterprises at the Departments of Labor, Education, and Energy 4
Federal Data Center Consolidation Initiative (FDCCI) Drivers and Opportunities Data Center Consolidation Drivers: Since 1998, the number of data centers increased from 432 to 2,094, while server utilization less than 30% typically* Significant increase in the quantities of data flowing into the enterprise; real time and offline analytics create significant challenges for most Agencies Increasingly mobile workforce, requiring computing capabilities wherever they work Business processes and applications must change to be able to accommodate increasingly unpredictable demand Data Center Consolidation Opportunities: Virtualization of servers can dramatically reduce data center floor space with conservative 20:1 ration of server consolidation Virtualized storage; variable capacity networks; self-protecting data; intelligent automated processing of data can handle enormous volumes of without exponentially increasing costs or resources Cloud solutions are internet based and support multiple devices with standard Web based interfaces Sharing and reusing applications & services across the enterprise; standardization; SOA, web enabled and composite applications Federal Cloud Computing Strategy, February 8, 2011 Vivek Kundra U.S. Chief Information Officer 5
Data Center Consolidation Key Cost Levers 30% 25% Data Center Key Cost Components per OS Operations (Manpower) Software Data Center Consolidation Cost Levers 30-40% data center costs savings may be realized across the Federal Government, through a combination of server virtualization, application rationalization, interagency colocation and moving apps to the cloud Operations (Manpower) 25% 10% 5% 5% Server Hardware Power Building & Facilities Network / ISP Software Server Hardware Power Building & Facilities Network / ISP 6
Barriers to Achieving Savings and Common Concerns Fiscal Challenges Lack of adequate / up front funding Establishing / terminating contracts Ability to transfer funds, color of money Apples to Apples cost comparison of options Technical obstacles Legacy applications, brittle architecture filled with undocumented dependencies and a patchwork of custom code Technical expertise Existing data centers simply inadequate Location dependencies (Print, domain controllers, latency issues ) Cultural and political challenges Change, Ownership, fear of unknowns, embracing new technologies Cloud new and relatively untested NIST / Open Cloud Computing Consortium still defining reference model A&A process proceeding but slow going 7
eglobaltech Data Center Consolidation Tools for Achieving Success Automation tools for autodiscovery & management Provisioning automation simplifies operations and drastically cuts costs Cut data center costs by over 20% by virtualization alone, reducing hardware and floor space As much as a 60:1 server consolidation ratio can be achieved, based on current average utilization rate Automate Virtualization Approaches to Reducing Data Center Costs Move to the Cloud Rationalize & Standardize Moving some applications to the cloud pushes significant IT infrastructure and operating costs out of Agency data centers Target applications to move to the cloud include web sites and Email Identify and eliminate redundant applications reducing software & app maintenance costs Increase sharing of infrastructure across the enterprise, leveraging virtualization Standardize to simplify and lower cost of operations 8
eglobaltech Data Center Consolidation High-Level Approach High-Level Methodology for Data Center Consolidation Dividing up the layers to focus on automation, rationalization/standardization and virtualization/cloud at each layer Eliminate redundant applications Consolidate common services (e.g. email) Assess Cloud (SaaS) alternatives Decommission idle & legacy servers Physical server consolidation & virtualization Assess Alternatives including Cloud (PaaS/IaaS ) and Outsourced Hosting Optimize Data center floor space, racks Rate & rank existing facilities Invest in the ones that make sense. Decommission the ones that don t 9
eglobaltech Data Center Consolidation Methodology & Tools Phase 1 Service and Asset Inventory Phase 2 Mapping & Analysis Phase 3 Rationalize & Model Phase 4 TCO Transition Planning Phase 5 Manage Execution METHOD: Identify and document HW/SW, and physical assets for all data centers Map infrastructure to applications / services to business processes and perform analysis Streamline application, service portfolios Model alternatives, capacity, costs Destination preps, sequencing, risk mitigation, execution planning Manage execution across the enterprise including HR, finance, IT and facilities TOOLS: Asset-discovery tools to capture existing assets Automated application mapping tools Capacity requirements, alternatives and cost projections Move sequencing & contingency plans, mitigation strategy Data center PMO tools manage HW, floor space, power and labor 10
Phase 1 Service and Asset Inventory Activities / Focus Areas Key Activities: Systems - types, utilization, OS s, Virtualization, storage capabilities Key Focus Areas: Data Center tier, size, power capacity, PUE utilization Facilities - Layout, expandability, availability of power, Cabling, fire suppression, security, age of systems Even height of ceiling and raised floor are important Network and storage infrastructure - cabling, port speeds Focus Areas Issues Cost Cutting Recommendations Under utilized servers Legacy Hardware Test / Dev Environments Web and Print Servers Poor utilization of resources Cannot be virtualized, security risk, high failure rates, high power consumption, compatibility issues (software, drivers and other hardware) Typically very dynamic and takes up a large percentage of servers in data centers Often the largest percentage of server counts within data centers, with least value Rationalize, decommission and virtualize Decommission or virtualize, establish standard minimum platforms Establish test/dev cloud in private or public clouds Virtualize or move to the cloud 11
Phase 2 Mapping & Analysis Activities / Focus Areas Key Activities: Mapping applications, services and systems Key Focus Areas: Catalogue services map to business Understand the business value Understand the real cost of delivering Map dependencies: understand what happens if you unplug it Understand minimum system requirements System average / peak loads, Storage, average / peak number of users, Bandwidth requirements (KB s per second, per week) Focus Areas Issues Cost Cutting Recommendations Gaps / Overlapping Applications Legacy Application Many applications doing the same function Outdated applications with legacy hardware requirements, with unsupported capabilities Rationalize applications find cloud alternatives Find cloud alternatives or upgrade to next level Manual Processes Processes to support applications manual Find opportunities for automation of software configuration 12
Phase 3 - Rationalize & Model Activities / Focus Areas Key Activities: Low hanging fruit Rationalize and streamline Application portfolio Eliminate redundancy Eliminate low use / high cost services Consolidate business line and agency ELA s Trim excess licenses from agreements Decommission legacy systems / applications Consolidate and standardize common services Validate systems/services provide desired business value Enterprise Architecture / Segment Architecture Identify and remove stovepipes Streamline Services Key Focus Areas: Focus Areas Issues Cost Cutting Recommendations Mix of a variety of OS and HW Platforms Lack of Data Center Automation Non-standard platforms cause limitations in consolidation add challenges to management and operations Manual processes for managing and operating systems Standardize and rationalize Find Cloud / Virtualization platforms Automation of processes for provisioning and management 13
Phase 4 Consolidation Planning Activities / Focus Areas Key Activities: Identify everything that s moving (Physical /logical) Dependencies, interdependencies, criticality Identify where it s going New facility/rack/slot location Develop move sequencing plan Key Focus Areas: Schedule of events, timeline, load & unload plan Identify required resources People, systems, logistics Identify gaps in new location resources Bandwidth, power, cooling, storage Develop risk mitigation plan What happens if something goes wrong / doesn t work What has to work if timeline slips, plan to unwind, go/no-go criteria Focus Areas Issues Cost Cutting Recommendations Lack of Broad Planning Lack of Governance Data Center specific planning with limited ability to coordinate across the enterprise Manual processes for managing and operating systems Establish enterprise-wide data center planning to coordinate Automation of processes for provisioning and management 14
Phase 5 Execution Activities / Focus Areas Key Activities: Prep new locations Expand services where required Bandwidth, network infrastructure, Storage Verify everything and everyone is in place Something as simple as mounting screws, power or patch cords can derail everything Execute the plan Measure results Efficiencies gained, cost savings, cost avoidance Update models as required Post execution Decommissioning, Update EA / Asset inventory, lessons learned, Key Focus Areas: Focus Areas Issues Cost Cutting Recommendations Inadequate Network or Storage capacity Manual Ops Processes Data centers that are receiving sites, may not have adequate network or storage for the new servers and applications Manual Operational processes can significantly delay and increase cost of transition of resources between data centers Ensure during execution to coordinate network resources Automation of processes for provisioning and management capabilities to streamline transitions 15
eglobaltech Data Center Consolidation Tool Suite DATA INPUTS Data Input from Multiple Formats (MS Excel, Access, CVS, etc.) eglobaltech Tools Performance Analysis Inventory Database Virtualization Analytics OUTPUTS FDCCI, Inventory & Metrics Reports Virtualization Cost, Energy / Performance Models Input Data from Existing Databases and Systems Custom and Ad Hoc Report Generation TCO Modeling User Friendly Graphical Interface Database (Access, SQL or Oracle databases) TCO Modeling 16
eglobaltech Data Center Modeling Tool The value of models Quickly identify high cost centers Explore what if s project over time Validate data collection Compare values to industry averages Identify and quantify shortfalls Space, power, cooling, storage Cost ROM, ROI, TCO Alternative analysis Apples to apples even if different cost structures Multiple frames of reference Agency, Business line, Data Center, project 17
eglobaltech Data Center Modeling Tool Components Facilities: Network and storage: Total sq.ft. Usable sq.ft. Total Racks Total Power White space power Load at PDU/UPS HVAC Tier Lease costs TI (MEP upgrade) costs # ISP lines Pipe size Network equipment Interfaces and speeds Average and peak utilization Total SAN / NAS / DAS Used storage Interfaces / speeds Servers and software: Personnel: # Servers Types of servers Average utilization # OS s Types of OS s # Virtual hosts OS Growth rate Automation, security, configuration management software costs # and type of personnel Average cost per FTE OS per FTE Server per FTE 18
eglobaltech Data Center Modeling Tools Evaluating Options Our Modeling Tool Can Assess Different Options for Data Center Consolidation ILLUSTRATIVE Based on our assessment building internal private clouds do not always yield optimal results, mainly due to investments in additional HQ/SW to manage, administer and provision IT services From an infrastructure standpoint, public cloud presents a compelling alternative, especially if server virtualization and cloud are combined Based on eglobaltech assessment of several Agency costs estimates 19
eglobaltech Data Center Modeling Tools Managing Resource Utilization Our Modeling Tool Can Provides a Clear Understanding of the Capacity for Each Data Center, Which is Critical to Planning Data Center Consolidaiton ILLUSTRATIVE Illustrative example of eglobaltech analysis of an Agency data center resource utilization forecasts We found that each data center is different and where identifying which resources may run out of capacity is also different. For example, one data center may have 50% of its floor space open but can not add anymore systems because of power limitations; while another may have plenty of cooling, but no floor space A more detailed analysis and inventory of data center capacity is required to truly optimize and consolidate across the an Agency and the Federal Government 20
FDCCI Working Groups & Focus Areas Working Groups IMPLEMENTATION GUIDES ALTERNATIVE OPTIONS Focus Areas Develop a technical and business implementation guide. Develop standard criteria by which all agencies can measure consolidation. Identify information to display in a dashboard available to the government and the public. Determine the information to publicly publish from an agency's final plans and consolidation efforts. This group will examine and brief the task force members on the existing government cloud initiatives. They will collaborate with the GSA cloud computing PMO to understand areas the task force can leverage to help the FDCCI program be more successful. TOTAL COST OF OWNERSHIP (standardize cost modeling) DATE CENTER GOV-WIDE MARKETPLACE (enable shared resource across Government) Review the various costs of ownership models that are in use today by agencies. The goal is to refine and establish one model with improved metrics that all agencies can leverage. Determine the need and viability of a multi-agency data center solution provided by government agencies. Develop and share with the task force members best practices as they relate to data center consolidation. 21
Contacts Nick Mistry Ph: (703) 585 8532 Nick.Mistry@eglobaltech.com Jeff Vandervort Ph: (757) 620 8015 Jeff.Vandervort@eglobaltech.com Thank You! 22