HOW TO PREPARE A DATA CENTER STRATEGY What you need to know to build a business case for Data Center sourcing Presented by: Gary Davis Principal PlanNet Consulting (714) 982-5886 www.plannet.net Steve Miano Principal PlanNet Consulting (714) 982-5820 Agenda ü ü ü ü ü ü ü What is a Data Center strategy? Ge;ng started Assessing the current state Developing a future state model Building a business case What your decision-­‐makers want to know Case studies Data Center Strategy A Data Center Strategy is usually iniBated for any or all of the following reasons: ü ü ü ü ü ü Data Center failures are impacBng the business Growth is expected to exceed power, cooling and/or space capaciBes Risk – real or perceived Lowering Cost of OperaBon Improve systems availability and disaster recovery Inform a decision to build, co-­‐locate or a combinaBon of approaches Ge;ng Started Define your objecBves ü Business drivers ü Technology imperaBves ü Desired outcome Assemble a Core Team ü IdenBfy/appoint a project sponsor • Sponsorship typically comes from Real Estate, FaciliBes or IT ü Typical make-­‐up of a Core Team • FaciliBes representaBve (VP, Director) • IT representaBves – CIO, IT Infrastructure VP, Network, user representaBve(s) • Procurement, sourcing and finance IdenBfy Stakeholders ü A Data Center strategy has many stakeholders ü It’s criBcal to engage to appropriate stakeholder resources to gain consensus and eliminate surprises ü Missing a key stakeholder is analogous to missing a Bre on your car Gather InformaBon Formalize the project with a Stakeholder kick-­‐off ü Introduce the iniBaBve and objecBves ü Layout Bmelines ü Set expectaBons Assessing the Current State Develop and issue a Comprehensive Request for InformaBon and meet with Stakeholders to gather informaBon ü Some of the informaBon is available electronically (e.g., CMDB, As-­‐Built drawings, energy bills, etc.) ü Some of the informaBon will be drawn out in Stakeholder meeBngs (e.g., applicaBon por\olio, financial models, etc.) RFI Categories ü ü ü ü ü ü ü ü ü ü Equipment/applicaBon inventory Data Center FaciliBes Real Estate OperaBons Network Business Systems & Services Business ConBnuity Finance Growth & Capacity Planning IT Strategic Plan Assessing the Current State Key Points ü Establish current demand load KW is the currency ü Determine historical demand growth (3-­‐5 years) ü Assess exisBng facility’s aaribute and constraints (e.g., Tier raBng, power, cooling, space, SPOCs) ü IdenBfy applicaBons and their Ber levels ü IdenBfy network constraints (e.g., latency sensiBve applicaBons) ü IdenBfy adjacency requirements/constraints ü IdenBfy key business & IT iniBaBves (e.g., new business ventures, merger/acquisiBon, virtualizaBon, etc.) ü Document outage history (many execuBves are unaware of outages) Outages….what outages? 11/18/2007 PRISM outage 3/12/2007 UPS Failure 6/30/2008 Spokane ATS Failure 16 hour outage 4,000 hours recovery effort 4/1/2007 12/31/2008 Failed UPS Transfer 7/1/2007 25 hour outage 6,300 hours recovery effort 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 1/1/2007 10/1/2009 10/26/2009 7/11/2007 Over Temp from Power Failure 5/16/2008 Power Upgrade Outage 23 hour outage 5,800 hours recovery effort 12/8/2008 Sprinkler Work Power Outage 6/25/2009 Fire Panel AC Outage Future State IdenBfy the future state scenario ü Establish operaBonal objecBves and service levels ü Establish availability targets • Tier RaBng • Geography ü Determine efficiency objecBves (PUE) ü Define sustainability goals ü Consider adaptability needs Plan for the Unplanned Establish the Future State Determine relevant sourcing models ü ü ü ü ü Co-­‐LocaBon Greenfield Tenant Improvement Retrofit ExisBng FaciliBes Cloud/Services Select markets to be evaluated ü Ideally, markets with co-­‐locaBon opBons, cheap power and free-­‐ cooling potenBal ü Determine business affinity to parBcular markets • ExisBng operaBons • Business drivers • ExisBng or future real estate or facility programs Establish Demand Forecast Align business growth with IT growth ü OpBmize IT deployment as part of the deployment of a new data center Capacity Demand Forecast 30,000 28,000 26,000 24,000 Capacity Demand -­‐ k Watts 22,000 20,000 18,000 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 FY09 FY10 FY11 FY12 Un-­‐optimized FY13 FY14 FY15 Managed Optimized FY16 FY17 FY18 Establish Demand Forecast ü Establish capacity threshold vs. demand forecast ü Depict historical growth with future planning ü IdenBfy strategies to miBgate capacity constraints Define Day 1/Day 2 Plan for incremental deployment based on demand forecast Site 1 -­‐ Production Buildout Phases 16,000 14,000 12,000 k Watts 10,000 *DAY 2: 7,900 kW, 22.5K SF 8,000 6,000 * DAY 1: 3,900 kW, 22.5K SF 4,000 2,000 0 FY09 FY10 FY11 FY12 FY13 Managed Forecast FY14 FY15 FY16 Deployed Capacity FY17 FY18 Market Analysis Qualitative Metrics use to disqualify markets Energy ü ü ü ü TransportaBon/Accessibility Rates vs. Nat’l average Sustainability Free cooling Reliability TelecommunicaBons ü Flight Availability ü Door to Door Time Vendor & Services Support ü Proximity to criBcal vendors Environmental Risk ü Diversity ü RelaBve costs ü Latency Labor ü Seismic/Volcanic ü Hurricane/Tidal ü Severe Weather IncenBves (very dependent on properBes) ü Enterprise Zones ü Sales Tax ü Energy rebate programs ü Availability ü RelaBve costs FaciliBes ü Control/Lease Terms QualitaBve Market Rankings What’s important to you? Keep it simple Building a Business Case Determine Capital Expenditures Determine OperaBonal Expenditures Develop a MulB-­‐Year Total Cost of Ownership for each model (5/10 yr) Compare TCO for each model Determine CapEx & OpEx CAPEX ü ConstrucBon & Sok Costs (Greenfield, TI) ü Cable Plant ü Network Electronics ü DepreciaBon ü MigraBon OPEX ü ü ü ü ü ü ü ü ü ü Lease costs Power Maintenance Facility Management Security IT OperaBons Staff TelecommunicaBons Services Sales Tax MigraBon Prepare TCO for each Scenario Co-Location Example E. Washington Co-­‐Lo (Phase-­‐In) 2011 Start Operating Costs Power Predictive Maintenance Rent IT Staff Facilities Staff Telecommunications Co-­‐lo Setup Costs DC Equipment Sales Tax Operating Costs Subtotal Capital Costs DC TI Racks / Cabling Network Electronics Migration Costs Capital Costs Subtotal Depreciation Cash Flow Total Year 1 Year 2 Year 3 Year 4 Year 5 5 year Total Year 6 $ 46 $ -­‐ $ 478 $ 90 $ -­‐ $ 828 $ 120 $ 128 $ 1,690 $ 96 $ -­‐ $ 915 $ 93 $ -­‐ $ 845 $ -­‐ $ 130 $ 2,079 $ 154 $ -­‐ $ 1,244 $ 96 $ -­‐ $ 861 $ -­‐ $ 133 $ 2,489 $ 166 $ -­‐ $ 1,269 $ 99 $ -­‐ $ 879 $ -­‐ $ 135 $ 2,548 $ 178 $ -­‐ $ 1,294 $ 102 $ -­‐ $ 896 $ -­‐ $ 138 $ 2,609 $ 641 $ -­‐ $ 5,201 $ 480 $ -­‐ $ 4,309 $ 120 $ 664 $ 11,414 $ -­‐ $ 1,112 $ 879 $ 1,624 $ 3,615 $ 1,958 $ 5,305 $ 3,649 $ -­‐ $ -­‐ $ 188 $ 915 $ 1,104 $ 1,288 $ 3,183 $ 3,367 $ -­‐ $ -­‐ $ 188 $ 634 $ 823 $ 1,296 $ 3,311 $ 3,785 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 410 $ 2,548 $ 2,958 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 410 $ 2,609 $ 3,019 $ -­‐ $ 1,112 $ 1,256 $ 3,174 $ 5,541 $ 5,363 $ 16,956 $ 16,777 Year 7 Year 8 Year 9 Year 10 10-­‐ Year Total $ 192 $ -­‐ $ 1,411 $ 105 $ -­‐ $ 914 $ 17 $ 141 $ 2,779 $ 205 $ -­‐ $ 1,439 $ 108 $ -­‐ $ 932 $ -­‐ $ 144 $ 2,828 $ 220 $ -­‐ $ 1,468 $ 111 $ -­‐ $ 951 $ -­‐ $ 144 $ 2,894 $ 237 $ -­‐ $ 1,497 $ 115 $ -­‐ $ 970 $ -­‐ $ 146 $ 2,965 $ 254 $ -­‐ $ 1,527 $ 118 $ -­‐ $ 990 $ -­‐ $ 146 $ 3,035 $ 1,749 $ -­‐ $ 12,542 $ 1,037 $ -­‐ $ 9,066 $ 137 $ 1,384 $ 25,916 $ -­‐ $ 127 $ 378 $ -­‐ $ 505 $ 328 $ 3,284 $ 3,107 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 290 $ 2,828 $ 3,119 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 94 $ 2,894 $ 2,988 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 94 $ 2,965 $ 3,059 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 94 $ 3,035 $ 3,129 $ -­‐ $ 1,240 $ 1,633 $ 3,174 $ 6,047 $ 6,262 $ 31,963 $ 32,178 Total Cost of OperaBon Tenant Improvement Example Seattle TI/Lease Operating Costs Power Predictive Maintenance Rent IT Staff Facilities Staff Telecommunications Other Costs DC Equipment Sales Tax Operating Costs Subtotal Capital Costs DC TI Racks / Cabling Network Electronics Migration Costs Capital Costs Subtotal Depreciation Total Year 1 Year 2 Year 3 Year 4 $ 297 $ 131 $ 462 $ 250 $ 147 $ 828 $ -­‐ $ 131 $ 2,246 $ 320 $ 221 $ 476 $ 258 $ 151 $ 845 $ -­‐ $ 134 $ 2,405 $ 345 $ 279 $ 490 $ 265 $ 155 $ 861 $ -­‐ $ 137 $ 2,533 $ 371 $ 233 $ 505 $ 273 $ 160 $ 879 $ -­‐ $ 139 $ 2,561 $ 400 $ 240 $ 520 $ 282 $ 165 $ 896 $ -­‐ $ 142 $ 2,644 $ 1,734 $ 1,104 $ 2,453 $ 1,328 $ 778 $ 4,309 $ -­‐ $ 683 $ 12,388 $ 6,078 $ 1,112 $ 1,157 $ -­‐ $ 8,347 $ 796 $ 3,041 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 796 $ 3,200 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 859 $ 3,392 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 796 $ 3,357 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 796 $ 3,440 $ 6,078 $ 1,112 $ 1,157 $ -­‐ $ 8,347 $ 4,041 $ 16,429 In thousands Year 5 5 year Total Year 6 Year 7 Year 8 Year 9 Year 10 10-­‐ Year Total $ 429 $ 247 $ 536 $ 290 $ 170 $ 914 $ -­‐ $ 145 $ 2,730 $ 460 $ 254 $ 552 $ 299 $ 175 $ 932 $ -­‐ $ 148 $ 2,820 $ 494 $ 262 $ 568 $ 308 $ 180 $ 951 $ -­‐ $ 148 $ 2,911 $ 530 $ 270 $ 585 $ 317 $ 186 $ 970 $ -­‐ $ 151 $ 3,009 $ 568 $ 278 $ 603 $ 326 $ 191 $ 990 $ -­‐ $ 151 $ 3,107 $ 4,216 $ 2,415 $ 5,296 $ 2,868 $ 1,680 $ 9,066 $ -­‐ $ 1,425 $ 26,966 $ 876 $ 127 $ 378 $ -­‐ $ 1,381 $ 716 $ 3,447 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 716 $ 3,537 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 557 $ 3,468 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 557 $ 3,566 $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ -­‐ $ 557 $ 3,664 $ 6,954 $ 1,240 $ 1,535 $ -­‐ $ 9,728 $ 7,145 $ 34,111 Compare Models Year 1 Year 2 Annual Operating Costs Years 1 -­‐10 (in 1000's) Year 3 Year 4 Year 5 5 year Total Dif $ Dif % 10 Year Total Dif $ Dif % As-­‐Is Optimized Owned Owned $ 2,929 $ 3,125 $ 3,461 $ 3,661 $ 3,878 $ 17,054 $ -­‐ $ 2,669 $ 2,869 $ 3,047 $ 3,232 $ 3,433 $ 15,251 $ (1,804) 0% -­‐11% $ 39,639 $ -­‐ $ 35,527 $ (4,112) 0% -­‐10% Market 1 Co-­‐Lo $ 3,145 $ 3,117 $ 3,281 $ 3,298 $ 3,383 $ 16,224 $ (830) -­‐5% $ 33,773 $ (5,866) -­‐15% Market 2 Co-­‐Lo $ 3,205 $ 3,188 $ 3,362 $ 3,391 $ 3,489 $ 16,635 $ (419) -­‐2% $ 34,896 $ (4,743) -­‐12% Market 3 TI / Lease $ 4,388 $ 3,653 $ 3,900 $ 3,882 $ 3,993 $ 19,816 $ 3,620 21% $ 44,423 $ 4,784 12% Market 4 $ 3,327 $ 3,400 $ 3,610 $ 3,677 $ 3,814 $ 17,828 $ 3,449 20% $ 38,203 $ (1,436) -­‐4% Market 5 TI / Lease $ 3,302 $ 3,497 $ 3,730 $ 3,717 $ 3,823 $ 18,068 $ 3,447 20% $ 38,071 $ (1,569) -­‐4% Market 6 20% $ 37,778 $ (1,861) -­‐5% Co-­‐Lo Co-­‐Lo $ 3,306 $ 3,368 $ 3,557 $ 3,622 $ 3,746 $ 17,599 $ 3,425 Scenario Analysis Annual Operating Expense Comparison $10,000 Annual O perting Expenses (1,000's) $9,000 Senario 1 $68.1M 1 0 Yr TOC $8,000 Scenario 2 $65.5M 1 0 Yr TOC $7,000 $6,000 $5,000 Scenario 3 (Current-­‐state with expansion) $58.7M 1 0 Yr TOC 2009 w/Stabalization $4,000 2009 $3,000 $2,000 $1,000 0 1 2 3 Scenario 2 4 Year 5 Scenario 1 6 7 Scenario 3 8 9 10 Case Studies Major Regional Bank Issues Results Analysis ü Aging primary data center in a high risk region ü Sub Tier 2 design ü History of outages ü Running out of capacity ü Facing major capital expenditures to retrofit the primary data center ü Lease expiring on Disaster Recovery site ü Assess deficiencies of exisBng faciliBes ü Determine amount of Bme before the DC reaches capacity ü Evaluate low-­‐risk markets ü Evaluate own vs. co-­‐lo Outcome ü Move primary DC to co-­‐lo ü Retrofit exisBng DC to Tier 2 ü Repurpose exisBng DC to sub-­‐ producBon and DR ü Improved availability with nominal increases in TOC Case Studies Global Technology Company Issues Results Analysis 40+ data centers worldwide Growing through acquisiBon Unsustainable operaBng costs Unmanageable data center footprint ü Business driver to consolidate operaBons ü Latency sensiBvity from global engineering applicaBons ü ü ü ü ü Assess operaBon efficiencies and costs ü Evaluate mulBple markets and corporate owned real estate ü Evaluate geo-­‐regional consolidaBon alternaBves ü Analyze enterprise architecture ü Evaluate own vs. co-­‐lo Outcome ü Geo-­‐regional consolidaBon (N.A, ENAME, PAC-­‐RIM) ü Deploy owned Tier 2+ ü MulB-­‐site recovery ü $100M+ 10-­‐year future cost avoidance Case Study Global Entertainment Company Issues Results } ü Significantly high energy costs ü High-­‐risk locaBon ü Planning for significant business technology iniBaBves ü Facing future capital expenditures to uplik primary data center Analysis ◦ Assess deficiencies of exisBng faciliBes ◦ Determine impact of new compuBng demand ◦ Evaluate low-­‐cost markets ◦ Evaluate own vs. co-­‐lo } Outcome ◦ Expand primary DC to co-­‐lo ◦ Leverage exisBng facility for business partner and DR ◦ Recommended phased deployment to opBmize ROI Case Study Regional Healthcare Company Issues Results } ü Aging primary data center in a high risk region ü Sub Tier 2 design ü History of outages ü Running out of capacity ü Facing major capital expenditures to retrofit the primary data center ü Affinity for in-­‐State deployment Analysis ◦ Assess deficiencies of exisBng faciliBes ◦ Assess miBgaBon strategies to buy Bme ◦ Evaluate regional markets ◦ Evaluate own vs. co-­‐lo } Outcome ◦ Move primary DC to co-­‐lo in a lower risk geography ◦ Repurpose exisBng DC to sub-­‐ producBon and DR ◦ Stabilize current environment ◦ Recommended best value alternaBve Thank you for your Bme!