Avoid Rolling the Dice: Solving the CIO s Data Center Dilemma with Colocation



Similar documents
How to Select a Colocation Provider Offering High Performance Computing

Scalability & Capacity Planning: Build vs. Buy

High Performance Computing (HPC) Solutions in High Density Data Centers

Brought to you by. Colocation Solutions BROUGHT TO YOU BY.

How to Obtain the Uptime, Security and Robust Connectivity Financial Services Firms

Interconnection for Financial Services

EXECUTIVE REPORT. Scalability & Capacity Planning: Build vs. Buy

CIOs Rising in Performance Needs

EXECUTIVE REPORT. Build vs Buy: Key Steps to Overcome Capital Constraints in the Data Center

The Sky for the Cloud : CyrusOne Data Centers. Fred Holloway Vice President, Partner Ecosystem CyrusOne, LLC

Federal Data Centers: The Build vs. Buy Decision

CYRUSONE COMPANY OVERVIEW BUILT FOR TOMORROW. READY TODAY.

CARFAX Selects Data Center for Connectivity & Redundancy

EXECUTIVE REPORT. The Data Center Build-or-Buy Decision: 6 Key Factors You Should Consider

EXECUTIVE REPORT. 4 Critical Steps Financial Firms Must Take for IT Uptime, Security, and Connectivity

How T-Systems Solved Its Data Center Challenges to Offer Dynamic Services to the U.S.

EXECUTIVE REPORT. Can Your Data Center Infrastructure Handle the Heat of High Density?

Powering Your Datacenter

The art of the possible.

Future IT Capacity Planning Depends On Flexibility

WHY EQUINIX WINS: A PRIMER FOR PARTNERS

Can Your Data Center Infrastructure Handle the Heat of High Density? 5 Ways to Test Your Data Center Infrastructure before Installing Blades

Ability to Build Hyper-Scale Data Centers at Hyper-Speed Helps Eliminate IT Expansion Risk and Uncertainty

STATEMENT OF. Dr. David McClure Associate Administrator Office of Citizen Services and Innovative Technologies General Services Administration

SCALABILITY & CAPACITY PLANNING: BUILD VS. BUY

CASE STUDY SOFTWARE. Zendesk s Cloud-Based Customer Ser vice Platform Lives in RagingWire s Mission Critical Data Centers

SOLUTION BRIEF CA Cloud Compass how do I know which applications and services to move to private, public and hybrid cloud? agility made possible

2011 EXECUTIVE REPORT. The Data Center Build-or-Buy Decision: 6 Key Factors You Should Consider

CASE STUDY. FuelQuest, Inc. Achieves 100% Uptime for their Cloud-Based Fuel Inventory and Tax Solutions at CyrusOne s Data Center

HOW TO SELECT A COLOCATION PROVIDER

Data Centre Outsourcing a Buyer s Guide

HOW TO PREPARE A DATA CENTER STRATEGY. What you need to know to build a business case for Data Center sourcing

Determining Total Cost of Ownership for Data Center and Network Room Infrastructure

CONSOLIDATING DATA CENTER SUPPORT:

2011 Australia Excellence Awards Datacenter Services Provider of the Year

CARFAX Selects Data Center for Web Hosting and Redundancy

Creating an IT Infrastructure that Adapts to Your Business PLAYBOOK

CASE STUDY. CyrusOne Helps City of Houston to Go Green and Reduce Energy Consumption through Data Center Consolidation

DATA CENTER COLOCATION

GET CLOUD EMPOWERED. SEE HOW THE CLOUD CAN TRANSFORM YOUR BUSINESS.

HOW MUCH ARE YOU REALLY PAYING FOR THAT kw OF POWER IN YOUR DATA CENTER?

New IT projects are difficult to finance due to the large upfront capital costs

Automated file management with IBM Active Cloud Engine

SERVICE OFFERING: DATA CENTER REAL ESTATE & SITE SELECTION SERVICES. Data Center Expertise Delivered with Integrity

Case Study In the last 80 years, Nationwide has grown from a small mutual auto

DATA CENTERS. VAZATA Headquarters: 6900 Dallas Parkway, Suite 800 Plano, TX VAZATA.COM

NTT Com Asia deploys NEC's Express 5800/ECO CENTER to become more competitive.

At Cavern Technologies, our goal is to meet and exceed our clients diverse requirements for IT infrastructure services.

Telephony & Connectivity to the Cloud

Accelerate server virtualization to lay the foundation for cloud

Moving Virtual Desktops to the Cloud

Case Study: Data Center Hosting for Information Technology Industry. Attachmate secures 100% uptime and seamless data center transition.

Evolved Capacity Management

Data Center Colocation Build vs. Buy

EXECUTIVE REPORT. Why Healthcare Providers Seek Out New Ways To Manage and Utilize Big Data

Executive Summary. Metro Capacity

Increasing Energ y Efficiency In Data Centers

Data Center & IT Infrastructure Optimization. Trends & Best Practices. Mickey Iqbal - IBM Distinguished Engineer. IBM Global Technology Services

Optimizing Power Distribution for High-Density Computing

GET CLOUD EMPOWERED. SEE HOW THE CLOUD CAN TRANSFORM YOUR BUSINESS.

Seven Practical Steps to Help You Run Your On-Premise Cloud Like a Business. Whitepaper

Cloud-Scale Datacenters

WWT View Point. Journey to the Private Cloud: Take the First Steps with FlexPod

GETTING THE MOST FROM THE CLOUD. A White Paper presented by

Colocation and Cloud Hosting Solutions. Build a strong and efficient physical infrastructure that delivers service reliability and profitability

Migration Planning and Execution:

Special Report Data Centers: Why the Great Demand for Data Centers?

Impact of Healthcare Regulations on the Data Center

What is a Data Center? Racks and Servers. At The Click Level. Equinix - Key Facts

Data Center Trend: Distributed Power in the White Space

Enabling IT Redundancy and Scalability for High-Availability Logistics Software

THE WORLD S MOST EFFICIENT DATA CENTER

CLOUD AND CARRIER- NEUTRALITY IN A COLOCATION DATA CENTRE

EX ECUT IV E ST RAT EG Y BR IE F. Server Design for Microsoft s Cloud Infrastructure. Cloud. Resources

RESILIENT PRODUCT LIFECYCLE MANAGEMENT STRATEGIES & SOLUTIONS FOR FUTURE-PROOFING PLM

high-performance computing so you can move your enterprise forward

The North American Data Centre Pricing 2014 to 2019 Report

Next Generation Cloud Services. White Paper

6Your Guide to a Superior

Audit of the Data Center Consolidation Initiative at NARA. OIG Draft Audit Report No May 10, 2012

A seismic shift in the market

Building vs Outsourcing. Data Center. A Business Proclamation by Leading CIO Fred Mapp. Presented by

Contact Centers in the Cloud: A Better Way to Source

ITM 6.1: To Colocate or Not Colocate

IBM Global Services September Disaster recovery strategies for internal recovery

Making IT serve the needs of BUSINESS.

Leveraging Cloud Computing for Disaster Recovery

High-performance. High-density. Highly committed. Built for tomorrow. Ready today.

Revitalising your Data Centre by Injecting Cloud Computing Attributes. Ricardo Lamas, Cloud Computing Consulting Architect IBM Australia

Data Center Network Evolution: Increase the Value of IT in Your Organization

Cloud Infrastructure Operational Excellence & Reliability

Case Study: D+M Group Cloud Consolidation: Migrating 15 physical data centers into three private clouds

Managing Data Center Growth Explore Your Options

(Examples of various colocation sites)

Data Center Optimization: California s Blueprint for Success

Brink s Chooses Data Center Provider to Consolidate Global IT Infrastructure

DCM 18.1 How to Prepare a Data Center Strategy

RagingWire Enterprise Solutions

Reducing Total Cost of Ownership through Outsourced Hosted Virtual Desktops

Executive Summary. Metro Capacity

Transcription:

EXECUTIVE REPORT Avoid Rolling the Dice: Solving the CIO s Data Center Dilemma with Colocation

Overview CIOs today face a litany of challenges and risks when it comes to meeting the goal of making the best possible IT investments for their organizations. Expensive, underutilized data centers, for example, can quickly become a financial drain for companies. This is true especially when they are looking to devote more of their IT budgets to core business innovation rather than to running and maintaining what they already have. According to Forbes Magazine s The Top 10 Strategic CIO Issues For 2013, 1 Far too many companies today find that they need to devote 70% or even 80% of their IT budget just to run and maintain what they ve already got, leaving as little as 20% for innovation. Forbes points to server sprawl, massively underutilized storage resources, unproductive data centers, labor-intensive integration requirements, and a near-endless list of strategic vendors as the biggest culprits. Far too many companies today find that they need to devote 70% or even 80% of their IT budget just to run and maintain what they ve already got, leaving as little as 20% for innovation. According to Forbes Magazine s The Top 10 Strategic CIO Issues For 2013 Additional IT investment risks the typical CIO faces in today s competitive business environment include the following: 1. Hardware Risk Which hardware platforms must be deployed to give the company the IT advantage it needs to drive revenue, increase productivity, and provide stronger analytical support to take advantage of emerging, big data demands? Is a high-power computing (HPC) solution needed? How much hardware will be virtualized? What are the forecasted power densities of the proposed hardware and how will storage be managed? These are all important questions that must be addressed before the time and investment in new IT infrastructure are made. 1 Forbes, The Top 10 Strategic CIO Issues For 2013, Sept 2012. http://www.forbes.com/sites/oracle/2012/09/28/the-top-10-strategic-cio-issuesfor-2013/ 2 20130805v2

2. 3. Software Risk - What software investments strengthen the organization s competitive advantage? How will mobile applications come into play and what will the SaaS (software as a service) roadmap look like? How much business will be done in the cloud over the next 5-10 years? Again, these are all critical challenges that CIOs must be able to effectively address in order to make the best possible IT investments. Business Risk What is the company s strategic roadmap? Will growth come via acquisition and/or will some units be divested? How nimble must the IT platform be in order to anticipate changing business strategies and growth through acquisitions? Scalability (both up and down) is a key factor in today s business environment, where organizations need nimble, flexible infrastructures that can adapt quickly and seamlessly. Technological Advancement Fuels Uncertainty With technology and business strategies evolving at an unprecedented pace, today s CIOs must be able to make good decisions despite the high amount of uncertainty they face every day. Consider these facts: Demand for digital information doubles every 18 months The typical hardware platform has a 3-year shelf life before becoming obsolete Roughly 1,000 new apps are being developed daily As if these pressures weren t enough for the CIO to grapple with, companies are moving forward with acquisition and growth strategies that significantly propel the number of possible permutations from an IT planning perspective. Rational data center strategies and leveraging new hardware and software technologies quickly are key issues for CIOs, who must ensure that data centers have the power and cooling necessary for tomorrow s IT applications. Adding fuel to the fire is the fact that the typical 50,000-square-foot, 5 Megawatt (MW) data center built by a Fortune 1000 company costs anywhere from $115 million to $200 million and has a useful life of about 20 to 30 years - if it s designed properly. The typical Fortune 500 CIO has a global data center portfolio of between three and six data centers to provide customers with high service levels, as measured by latency, and a resilient business continuity plan in case of unforeseen data center problems. The CIO must consider spending $115 million (at minimum) on a data center asset that will last for 20-30 years and house hundreds of millions of dollars in hardware and software assets, each of which has a shelf life of three to four years. Given that enterprise companies do not often build data centers, it generally takes two to three years to complete the construction on a data center project at a much higher expense than anticipated. 3 20130805v2

Colocation Can Address Many Challenges Most of these challenges can be addressed through colocation, or the outsourcing of data center services. More and more Fortune 1000 companies are turning to colocation providers to deliver buildings, power redundancy, cooling redundancy, physical security, fire suppression, and access to telecommunications carriers to a third-party provider that can take over the reigns from concept to completion and beyond. CyrusOne builds many data centers each year and can typically complete a new data center project from within 14 weeks to six months, depending on whether it is an existing data center pod fit out or a new, ground-up facility. All of this is done at a cost that s roughly 70 percent less than what an enterprise company would pay. As shown in the chart below, the typical enterprise will spend approximately $115 million building a 5 MW facility, which is about $80 million more than what CyrusOne would pay to build a similar structure. Cost to build By Fortune 1000 CyrusOne Cost to Build Savings Price Per Mega Watt $23 Million $7 Million Data Center Capacity 5 Mega Watts 5 Mega Watts Total Cost $115 Million $35 Million $80M or 70% 4 20130805v2

The Real Cost of Building a Data Center The decision to undertake a data center construction project is risky, given that such projects are new terrain for the typical enterprise company. The forecasting uncertainty of the IT demands of the company make building a data center even riskier because it literally requires a highly predictive crystal ball to ensure the investment will be sized properly. This is an impossible investment for the CIO to make correctly. Determining how to optimize a 20-year data center asset investment is another challenge. Assuming that the infrastructure environment is STATIC and allows for an optimized per-kilowatt (KW) of square foot of white space, for example, is the wrong approach in an environment of rapid technological advancement. Trying to determine software, hardware, or communication trends over the next three years is difficult enough; attempting to stretch that time period to 10-30 years is impossible. To get around this issue, most CIOs use a simple thumb in the air approach based on general assumptions like, Let s assume our existing power and square foot print grows by 5 percent a year, and then project forward for 15 years. They then use that as the basis for estimating their data center needs, which then serves as the starting point for establishing the size of the data center. This method is not only highly ineffective in today s evolving technology environment, but it can also cost the organization valuable time and financial investment over the long term. The forecasting uncertainty of the IT demands of the company make building a data center even riskier because it literally requires a highly predictive crystal ball to ensure the investment will be sized properly. This is an impossible investment for the CIO to make correctly. 5 20130805v2

Before You Know It, Your Data Center Could Be Obsolete As illustrated in the chart below, the minute a CIO commits to building a data center of a certain size, that facility is either too small or too large. The CIO made the wrong bet: if the facility is too large, then the cost per utilized KW or square foot, is significantly higher than what was assumed in the STATIC model. On the other hand, if the facility is too small then the CIO will have to allocate more time and expense to designing and building a new facility. Forecasting errors comprise one of the riskier propositions of building a new data center facility. In the chart below, CyrusOne s cost per MW is fixed at $7 million per MW, compared to the $23 million per MW cost that the typical enterprise company would build at. For the enterprise to obtain even the $23 million per megawatt level it would need to ensure that the data center demand is 5 MWs. The reality however, is due to the forecasting uncertainty around IT as well as how long it takes to migrate and fill up data centers, the actual cost per MW will be substantially higher. For this exercise, assume the enterprise customer just built a data center and starts utilizing 1 megawatt of capacity and then increments that utilization by 1 MW each year. As you can see from the chart, data centers are very expensive fixed cost investments and the actual price per utilized MW is substantially higher than what the enterprise CIO expected. At the beginning of the investment, the enterprise CIO will actually be paying the full $115 million cost for that first megawatt of use. As the MW utilization increases overtime, the cost per MW decreases and eventually reaches the $23 million per MW. 120 100 80 60 40 20 Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7 Year 8 Customer Cost per MW with 1MW demand but incr. by 1MW Customer Cost per MW with immediate 5MW demand Customer Cost from CyrusOne per MW 6 20130805v2

The cost of data center forecasting variability can be enormous and long lasting. If the Fortune 500 CIO builds a $115 million data center that is only 50% utilized because more of the company s IT computing was done in the cloud, for example, or more of the IT footprint was virtualized, then the company s realized price per KW or price-per-foot will be twice as much as originally predicted. The outcome gets even uglier if you consider that even if the enterprise company were to perfectly predict its IT demand, the green line reveals that CyrusOne s cost to the company would still be 70 percent lower than the organization s best-case scenario. Knowing they can t efficiently manage the planning uncertainty and the residual risk associated with building such an expensive, long-lived asset, more and more Fortune 500 CIOs and CFOs are reconsidering significant capital investment management. Why Wholesale Data Centers Don t Work The above analysis reveals the high cost of investment mistakes due to the planning uncertainties surrounding standalone data center construction. The problem of the STATIC analysis is no less of an issue when choosing to go with the standard wholesale data center provider. Wholesale providers lease a 1 MW, 10,000-square-foot data center pod to a customer for 10 years. The lease is generally structured as a base price for the infrastructure and increases every year by 3 percent on average. A wholesale customer pays separately for the power the 10,000-square-foot pod consumes, plus a portion of the building s operating expenses, including common area maintenance, taxes, depreciation, etc. These pass-through expenses can drive up the actual price charged by a wholesale provider by nearly 50 percent in some cases. In the wholesale product, the data center provider offers a capital investment solution that enables the customer to purchase 1 MW of data center capacity. The wholesale operator does not provide any services, which means the wholesale data center customer must staff the facility in order to manage the infrastructure and IT equipment including server reboots, simple tape rotations, and other types of services. The wholesale operator acts like a typical landlord and has the Just pay me the rent and I ll see you when the lease expires, mentality. From a financial perspective, the wholesale solution forces the inherent planning uncertainty associated with changing software, hardware, and business strategy trends that was introduced at the beginning of this report. To be optimal for the enterprise CIO, he or she must commit to using 1 MW of capacity for 10 years. If the CIO s IT demands change due to increased use of virtualization or because more applications are hosted in the cloud, the organization is obligated to pay for a 10-year contract. Assuming a 1 MW wholesale solution costs $2 million annually, a 20-year contract would cost the company $20 million excluding the 3 percent inflation increases. 7 20130805v2

Contrast This Scenario with CyrusOne s Flexible Colocation Solution Unlike a wholesale operator, CyrusOne sells exactly what the customer needs today in terms of space and power on a flexible basis. As the organization s IT requirements grow (or shrinks), the solution can be quickly scaled up or down to accommodate those needs. Contract flexibility is invaluable for CIOs trying to balance an uncertain business environment with budget issues and other constraints. What About Modular Data Centers in a Box? Modular data center products offer quick, in-the-box, data center solutions that can be delivered to a customer s location within 16 to 24 weeks. The organization decides exactly how much power it needs and then the modular manufacturer delivers the necessary infrastructure to support that demand level. While attractive on some levels, the modular data center presents several challenges for CIOs. This option doesn t address the planning uncertainty. Such products can be deployed quickly, but it doesn t alleviate the challenge of purchasing a modular solution that is built exactly as needed, rather than too small or too large to meet the enterprise CIO s needs. The actual cost of a modular solution is substantially higher than the colocation option. For example, CyrusOne s Massively Modular approach leverages economies of scale in order to drive down the cost per MW of data center space to a point that cannot be obtained in modular solutions. CyrusOne s Dallas facility, for instance, includes 700,000 square feet of capacity about 700 times larger than the typical pod solution. A cost-efficient modular solution would literally be working against the laws of economies of scale. The fact is, one million 1-foot-square boxes are not fundamentally more efficient than a single, 1 million square foot box. If the former were true, then every home in the U.S. would be running on its own personal electrical generator. Many companies don t realize data center in the box solutions lock them into today s physical server technology designs. What if the size of the standard blade server or rack changes? What will the Fortune 500 CIO do then? This is a real and credible risk companies don t always think through when selecting typical modular data centers. 8 20130805v2

Modular data center products offer quick, in-the-box, data center solutions that can be delivered to a customer s location within 16 to 24 weeks. The organization decides exactly how much power it needs and then the modular manufacturer delivers the necessary infrastructure to support that demand level. CyrusOne is building some of the largest and most energy efficient Massively Modular data centers today using a design philosophy leveraging massive economies of scale to achieve the lowest cost per MW in the industry. For example, the CyrusOne Phoenix facility will have 1 million square feet of 2.25 MW raised data center floor the largest in the region. Enterprise CIOs leveraging CyrusOne s data centers not only benefit from the best pricing in the industry but, due to the building scale and large number of other Fortune 500 customers in the facilities, they also gain from CyrusOne s better asset utilization and ability to pinpoint and manage forecasting uncertainties. CyrusOne offers best in class data center facilities that take the risk out of data center decisions. 9 20130805v2

About CyrusOne With over two dozen data centers across the globe, CyrusOne helps many of the world s largest global businesses including 9 of the global Fortune 20 companies and over 130 of the Fortune 1000 and companies of all sizes take advantage of the latest data center technology and realize top operational efficiencies through: Flexible design Scalable, customized data center solutions engineered with Massively Modular data center technology to align with your business needs Personalized service High touch customer service delivered by data center experts Full transparency Full transparency in communication, management, and service delivery High reliability Excellent availability using state of the art technology backed by 100 percent service level agreements (SLA) CyrusOne National IX Offers low-cost metro connectivity and city-to-city transport in an ever growing number of cities across the US. About the Author Gary Wojtaszek is the president & chief executive officer of CyrusOne and a member of the board of directors. Having previously served as Cincinnati Bell s Chief Financial Officer, Mr. Wojtaszek brings to the board of directors critical knowledge and understanding of the data center colocation business coupled with an indepth understanding of the company s capital structure. Prior to joining Cincinnati Bell, Gary served as the senior vice president, treasurer, and chief accounting officer for Laureate Education Incorporated in Baltimore, Md., where he was responsible for global controller and treasurer functions. Prior to working at Laureate Education, he was the vice president of finance and principal accounting officer for Agere Systems, Inc., a leading manufacturer of integrated circuits used in telecommunications and networking equipment, hard-disk drives, and other devices. While with Agere, Gary also held the positions of assistant treasurer, finance director and corporate controller. His past experience also includes treasury and corporate finance assignments with Delphi Automotive Systems and General Motors. Gary holds a bachelor s degree in economics and history from Rutgers University and a master s degree in finance and accounting from Columbia University. 10 20130805v2