Alaska Green Data Centers

Published on May 2016 | Categories: Types, Business/Law | Downloads: 30 | Comments: 0 | Views: 358
of 41
Download PDF   Embed   Report

Student Paper Submitted For ITM 6000 ProjectAuthor: Jason C. [email protected]: This paper reviewed secondary research and case studies to examine the economic and technical feasibility of building data centers in Alaska by demonstrating how the combined use of best practices, green technology, and the unique collection of renewable energy resources found in Alaska can be leveraged to improve profit and while lessening the environmental impact of growing data storage requirements. The paper also used India as a basis of comparison to demonstrate how Alaskan offshore outsourcing offers an improved quality of service, better security, and a more advantageous position in the current domestic political climate than does foreign offshore outsourcing.

Comments

Content

Data Centers Running Head: ALASKA GREEN DATA CENTERS

1

Feasibility of using Intraregional Green Energy Integration to Power Interregional Data Centers in Alaska Jason C. Murray Webster University

Data Centers Table of Contents

2

Title Page Table of Contents Chapter 1 Situation Analysis Statement of the Problem Premise Definitions Limitations Chapter 2 Review of Solutions to the problem with electricity Methods for Improving PUE Methods for Improving ERF Methods for decreasing nuclear/fossil fuel dependence Review of issues with datacenter outsourcing Security Loss of US Jobs IP Latency Suitability/Benefits of Alaska Location Security Telecommunications Infrastructure Assessment of Human Capital The Impact of Native Corporations

Page 1 Page 2 Page 4 Page 4 Page 7 Page 7 Page 7 Page 8 Page 9 Page 9 Page 12 Page 12 Page12 Page 14 Page 14 Page 17 Page 18 Page 19 Page 19 Page 20 Page 22 Page 23

Data Centers Tax Climate Climate Assessment Green Energy Potential Risks Chapter 3 Proposal and Recommendations Financial Analysis Financial Comparison Payback Analysis Chapter 4 Conclusions References Appendix A: System Request Sheet Appendix B: Project Definition Sheet Appendix C: Abstract Page 23 Page 24 Page 24 Page 24 Page 26 Page 26 Page 27 Page 29 Page 29 Page 29 Page 29 Page 31 Page 36 Page 37 Page 41

3

Data Centers Chapter 1 Situational Analysis: Modern society is rapidly becoming completely dependent upon internet-based technology. Improvements to broadband internet capability and availability, along with increased public confidence in internet based services, are changing the way we communicate, collaborate, and consume. Data centers are supporting this change by providing for the large volumes of data storage and processing

4

power that is required. Moreover, the move to cloud computing and internet-based commerce, as well as the increasing role of social media and real-time entertainment services in our daily lives are creating an even greater need for expanded data center services.

What exactly do data centers support?

Software as a Service (Saas) “is a model of software deployment where an application is hosted as a service provided to customers over the internet” (Shelly and Rosenblatt, 2012). It includes many popular cloud-based services such as online data backup/storage, application delivery, business database support, and hosted business services such as third-party billing and customer relationship management. SaaS continues to grow because there is no need to purchase, support, and maintain IT equipment or hire a robust IT staff. From a financial perspective, the benefits of SaaS make sense for large businesses looking to cut costs, small businesses that do not need and cannot afford a full time IT staff, and individual consumers. As the use of the SaaS model continues to grow, so does the need for more data centers to provide hosting services that support the SaaS market.

At the same time, internet based Business to Consumer (B2C) and Business to Business (B2B) commerce continues to grow. Online businesses are thriving, and by one estimate “the internet accounts for, on average, 3.4% of GDP across the large economies that make up 70% of global GDP” (Manyika & Roxburgh, 2011). The same study further estimated that the “Internet accounted for 21% of GDP

Data Centers growth in mature economies over the past 5 years” (Manyika & Roxburgh, 2011). Online business requires massive data storage and equally massive transaction processing capability. Additionally, historical data must be available, secure and easily accessible to both comply with potential legal requirements (which vary depending on the industry) and to support the use of data mining and business intelligence tools that allow a business to analyze and optimize its business performance. Computing and storage requirements to support and sustain the robust online business community are provided by data centers.

5

Finally, the expanding role and impact of social media, real time entertainment, and other Web 2.0 platforms in our daily lives cannot be understated. There are literally hundreds of millions of socialmedia accounts that collectively require significant storage and processing capability that is provided by data centers. Facebook alone has over 800 million worldwide users (Manyika & Roxburgh, 2011). The popular online role playing game World of Warcraft is reported to have over 10 million registered accounts (Karmali, 2012). Finally, the latest Global Internet Phenomenon Report (Sandvine, 2011) estimated that 53% of all US bandwidth in 2011 was used for real-time entertainment with Netflix streaming movie delivery alone using 29% of the real-time entertainment share. All of these capabilities are completely dependent upon data center services across the country to regionally stage and deliver data.

The major problem with current data centers: energy usage.

Unfortunately, data centers require large amounts of electricity to provide the abovementioned services. Adding the needed additional data center capability to support increasing internet-based technology usage will require ever-increasing quantities of electricity. At current, most of the required electricity used to operate existing data centers is generated from a combination of non-renewable

Data Centers environmentally damaging fossil fuels (coal and natural gas) and aging, potentially dangerous nuclear power.

6

The most significant source of electricity usage within a data center is climate regulation to create an environment suitable for the sustained operation of IT equipment. One study estimated that the combined use of energy for servers, other computers, UPS, power supplies, and fans accounted for 56% of data center electricity usage while air conditioning alone accounted for 44% of electricity usage (Tsuda, Tano, & Ichido, 2010). The same study further estimated that a 2000-rack commercial data center requires 87.6 million kilowatt hours of electricity annually just to operate the air conditioning (Tsuda, Tano, & Ichido, 2010). Using an estimate for the 12-month rolling US national average cost of $.0678 USD per kilowatt hour for industrial customers (US Energy Information Administration, 2012) a 2000-rack data center can be estimated to spend over $5.9 million dollars a year on air-conditioning alone. The same 2000-rack data center can be estimated to use almost 200 million kilowatt hours of electricity per year for all energy requirements (including air conditioning) at a cost of over $13.5 million USD per year.

Several green technologies currently exist that have the potential to significantly decrease the requirements for using fossil fuel electricity or nuclear power to provide air-conditioning for data centers. At the same time, recent advancements in green energy may allow for overall reductions in data center reliance on both fossil-fuel and nuclear energy altogether.

The second major problem with current data centers: outsourcing.

As stated, many existing data centers are not terribly energy efficient. Moreover, environmental regulations in the United States increase the cost of electricity provided through fossil fuel and nuclear energy sources in the form of energy taxes. These energy tax costs are subsequently passed on to

Data Centers customers. While these taxes are used for a variety of worthy causes to prevent and repair damage

7

done to our fragile eco-system, they still affect the bottom line for business customers. As a result, data center needs are now being outsourced offshore to providers in countries with less environmental regulation and cheaper energy that can provide services to anywhere in the world via the fiber optic link around the globe (FLAG).

Offshore data centers present several problems. First, data stored outside the United States is less secure and vulnerable to exploitation and theft than is data stored within the United States. Second, data center outsourcing reduces the number of high quality tech sector jobs available within the United States. Third, offshore data centers narrow the local, state, or federal tax base because they pay no taxes within the United States nor do their employees pay any taxes within the United States.

Statement of the Problem:

Are green technology data centers located in Alaska economically and technically feasible at this time?

Premise:

Building Data Centers in Alaska that are powered by green technology will result in reduced energy usage, lower operating costs, decreases in the release of environmental contaminants, and no change in the quality of service when compared to data center operations in more temperate areas of the continental United States. Finally, data centers in Alaska will be able to offer an improved quality of service and much better security when compared to offshore data centers in India.

Definitions and Key Concepts:

Data Centers -

8

Economical Feasibility: A proposal is considered economically feasible when the projected return on investment for a given proposal is higher than the projected return on investment using existing methods. This will be evaluated in this study using payback analysis.

-

Technical Feasibility: A proposal will be considered technically feasible when evidence is presented that establishes that all technologies upon which the proposal is dependent are tested, reliable, and appropriate to the proposal. This will be evaluated in this study by examining case studies of existing technology.

-

Power Usage Efficiency (PUE) is defined as Total Facility Power*/Total IT equipment power. PUE provides a comparison between how much power is used for actual IT equipment and how much power is used to power everything else (*measured at the data center utility meter). An ideal PUE of 1.0 means that all the energy used at a datacenter is consumed only by IT equipment and is considered to be the industry ideal. (Green Grid, 2012).

-

Energy Recovery Factor (ERF) is the ratio of energy saved by re-using waste heat. It is calculated as reuse energy (heat)/ total energy (heat) required. A facility that recovers no energy from the re-use of waste heat is considered to have an ERF of 0 (0/100 or 0%). A facility that provides 100% of heat from reuse energy is considered to have an ERF of 1 (100/100 or 100%) (Green Grid, 2012).

-

Intra-Regional Energy Integration: The use of energy resources in one region to provide for the energy requirements of data resources in another region.

Limitations: This is an initial feasibility research study for senior management using publically available data and existing research. It should be approximately 20-30 pages in length and address the problem statements and strategic goals listed in the project definition sheet.

Data Centers -

9

No more than $100 will be used to purchase necessary research references (studies, journal articles, etc). Funds for additional research or a full-fledged project will be provided at a later date dependent upon senior management’s assessment of the initial feasibility study.

-

Data center electricity usage estimates that are used to process economic calculations for determining economic feasibility will be based solely on numbers reported in existing research.

-

Estimates for how much energy might be needed to offset heating requirements for Alaskan-based data centers during winter months is not available. Chapter 2

Review of potential electricity solution case studies – methods for reducing PUE, improving ERF, and decreasing nuclear/fossil fuel energy dependence. Methods for reducing PUE: Most current industry efforts to reduce electricity usage in data-centers revolve around strategies to lower PUE. Ideal Power Usage Effectiveness (PUE) for a data center is 1.0 which represents the point at which power in a data center is used exclusively by IT equipment. A 2012 survey of over 1100 data centers worldwide found that the industry average is a PUE between 1.8 and 1.89 with 9% of respondents reporting a PUE of greater than 2.5 on the high end and 3% of respondents reporting a PUE of 1.3 or less on the low end. (Stansberry & Kudritzki, 2012). Some case studies of interest when considering how to lower PUE include the National Renewable Energy Laboratory (NREL) Research Support Facility (RSF), Google Data Center best practices, and the University of Electro-communication Systems Graduate School of Information System’s Hokkaido data center in Japan.

National Renewable Energy Laboratory (NREL) green data center – Golden Colorado:

Data Centers NREL conducted an 11-month study of an existing legacy data center and a new green data center (NREL, 2011). The result was an average PUE of 2.28 for the legacy data center and an average PUE of 1.16 for the green data center. NREL was able to achieve this low PUE primarily by implementing a natural air evaporative cooling system (which took advantage of low temperature air during Colorado winters) that was supplemented with traditional chilled water cooling during the warm months when

10

natural air temperatures were too high to be used for evaporative cooling. NREL also took measures to increase the efficiency of cooling within the data center by using an industry best-practice cold aisle/hot aisle configuration for IT equipment and then sealing the hot aisle to capture the exhaust and prevent it from mixing with the cold air supply. Cold air was pulled in from the front of the racks and hot air exhausted into sealed hot aisles through the back of the racks. Next, NREL used virtualization to reduce the number of physical servers – typically consolidating 20 physical servers into one. Finally, NREL applied a cable management strategy, blanking panels (for open rack spaces on the cool side of a rack), and equipment cowlings (for equipment that did not match rack depth) in order to maximize efficient cool air flow from the front to the back of racks. The ultimate result was an estimated savings of over $82,000 a year using NRELs utility rate of $0.057/kWh. Google Data Centers: Internet Search provider Google, Inc. claims an aggregate PUE of 1.13 across all its data centers and has been able to lower the PUE of one data center to as low as 1.06 (Google, 2012). Google has done this by using a variety of techniques at its various data centers around the world, but overall recommends five specific strategies for improving efficiency: (1) Measure PUE as accurately as possible to establish benchmarks that can be used to measure improvements in energy efficiency.

Data Centers (2) Manage air-flow through well-designed containment to minimize hot and cold air mixing. This includes a hot-air capture strategy and the use of blanks and cowlings to ensure that cold air flows

11

across equipment and is exhausted into the hot aisle where it can be captured. Consider using thermal modeling to identify hot spots and optimally design a data center conducive to efficient cooling. Finally, Google has taken hot-air/cold air capture and separation to a new level by reconditioning hot air exhaust through the use of water cooled “heat huts” and returning cool air to the ambient air in its data centers. (3) Adjust the thermostat in the cold aisle in accordance with American Society of Heating, Refrigerating and Air Conditioning Engineers (ASHRAE) recommendations. Google asserts that there is no decrease in performance with temperatures in the cold aisle set to as high as 80 degrees Fahrenheit. (4) Use natural air economizers to take advantage of evaporative cooling and the natural climate. Google takes this one step further at its Gulf of Finland data center where it uses cold sea water to provide liquid cooling for its IT equipment. Google also uses filtered waste water from the city sewer at its Belgium facility to help cool that data center as well. (5) Optimize power distribution. Google uses “highly efficient voltage regulator modules to ensure that most of the power goes to the components that do the actual computing work” (Google, 2012). Additionally they have “cut out two of the AC/DC conversion stages by putting back-up batteries directly on the server racks” (Google, 2012). Google estimates that by optimizing power distribution they have "achieved an annual savings of over 500 kWh per server—or 25%—over a typical system” (Google, 2012). Hokkaido, Japan Green Data Center:

Data Centers

12

Data from a Swedish snow-mound powered air-conditioning project for a hospital showed that densely packed snow mounds can be effectively preserved in hot summer months by packing the snow in a pit with as little as two-tenths of a meter (about six inches) of sawdust as a covering. The preserved snow was later melted in a controlled manner using waste heat from the hospital to provide a source of cool water for an evaporative cooling system that provided the hospital with AC during warm summer months (Skogsburg & Nordell, 2001). The snow-mound cooling technology employed in the Swedish study was taken one step further and applied in conjunction with natural air cooling systems at the Hokkaido data center in northern Japan. The Hokkaido data center uses fiber optic connectivity within the country of Japan to provide reliable and fast data services to Tokyo-based customers from a data center located hundreds of miles away in the northern cold-weather region of Hokkaido. The cooler weather in Hokkaido allows for natural air cooling when the temperature is below 60 degrees F, but more importantly, annual snowfall allows for cooling using snow mound storage technology (as part of an evaporative cooling system) when the temperature is above 60 degrees F. This combination of natural air cooling and the cooling power of stored snow allowed for the sustained year-round operation of a data center in Hokkaido with no energy used for the air-conditioning system even when the outside temperature in Hokkaido was at the local peak temperature of 85 degrees Fahrenheit (Tsuda, Tano, & Ichido, 2010). Methods for improving ERF: The second method to reduce data center electricity usage is to improve ERF. The Energy Recovery Factor (ERF) is the ratio of energy saved by re-using waste heat. An ERF of 100% indicates that a facility uses no additional energy to provide necessary heating for the facility. The NRL RSF in the aforementioned study was able to maximize ERF for nearby office facilities by redirecting waste heat from sealed hot aisles into the heating system of the nearby offices. This is particularly important when

Data Centers considering the cost of overhead energy required to heat offices occupied by data-center workers in

13

cold weather regions or even to warm IT equipment itself in extreme cold weather. Finally, waste heat is required to melt snow that generates the cold water used in snow mound technology air-conditioning systems. Methods for decreasing nuclear/fossil fuel dependence: The third method for decreasing data center electricity usage is not a data center-specific method for reducing electricity usage, but rather is concerned with using alternative energy sources to provide the electricity needed to run a data center. The idea is to generate low-cost, environmentally-sustainable energy from sources other than dangerous nuclear and high-polluting fossil fuel methods. Potential alternative energy solutions of note in Alaska include wind, hydro-electric, and geo-thermal energy. The feasibility of the aforementioned alternative energy solutions has been demonstrated by several energy programs of note including: The Fire Island Wind Project: Fire Island is in the Cook Inlet, just south of Anchorage Alaska, and is the location of a new wind farm being built by CIRI, an Alaska Native corporation representing a conglomeration of Native peoples in South-Central Alaska. When completed this fall, the 11-turbine wind farm is expected to provide the city of Anchorage with an additional 17.6 MW of power generation capability (approximately 4% of its electricity requirements - about 4000 households) at a competitive cost of only $.097 USD Per Kilowatt (CIRI, 2012). Kodiak Wind Project: In 2008, the city of Kodiak Alaska decreased it diesel-fuel generated electrical dependence from 20% to 11% (saving over 530,000 Gallons of diesel fuel per year) with the construction of three wind turbines on Pillar Mountain (Scott, 2010). Snettisham Hydroelectric Dam: Snettisham is a small town located approximately 28 miles from Juneau, Alaska and the location of a hydroelectric dam that takes advantage of two local lakes created from

Data Centers

14

annual snow run-off. The Snettisham dam currently provides approximately 80% of the power required in the Juneau-Douglas area municipality and sells excess electrical generation to the nearby Greens Creek Mine and to Princess Cruise Ships (AIDEA, 2011). Chena Hot Springs Geothermal Project: Chena Hot Springs is a small resort village located approximately 45-minutes north of Fairbanks, Alaska. In 2006 the Chena power company installed two small scale geothermal plants (at a cost of about USD $2 Million) using adapted HVAC technology to save on part fabrication costs. Chena now generates 400 kilowatt hours of power (with plans to install another 280 KW of power) that provides the local community and the resort (to include an indoor ice museum kept at negative 20 degrees Fahrenheit) with 100% of needed electricity at a cost of $0.06 USD per kilowatt hour. Excess generation is sold to the city of Fairbanks at a profit. Prior to the establishment of the geothermal plant power was provided via a diesel generator at a cost of over $0.30 per kilowatt hour (Holdmann, 2007). Initial capital costs of building a geothermal plant are estimated at about $3000 to $3900 per kilowatt hour of electricity for a 100 Megawatt plant with capital costs for smaller plants expected to be a bit higher. (Alaska Center for Energy and Power, 2012).

Review of issues with datacenter outsourcing: Security breaches, loss of US jobs, and IP latency:

Security:

The most significant threat posed by the outsourcing of data center services to offshore locations is data security. Offshore datacenters have several security flaws including physical security, employee suitability, and the issues with the application of US law (with respect to intellectual property rights, contracts, criminal prosecution, and US data security regulations).

The most basic precaution to ensure data security is physical security. If a security threat has physical access to your data, your data is no longer secure. In practice, this means that you should use a variety

Data Centers of physical security methods to prevent unauthorized physical access to your data, including fences,

15

gates, doors, locks, guards, mantraps, alarms, cameras, and identity confirmation when granting access. It also means that you should carefully vet anyone to whom you give access to your data and carefully monitor their activities for any sign that they have threatened your data security. Finally, you should carefully classify your data according to sensitivity levels and conduct regular audits to ensure that sensitive data is only accessed by authorized vetted personnel that have a genuine need to access that sensitive data. It is possible that these things can be done effectively when the data center is located in a foreign country, but not with as much certainty as with data located in the United States, which is subject to US litigation. Let’s consider some of the threats that would exist in a hypothetical decision to outsource data services to India:

India has a well-developed IT work force and is by far the most popular choice for US outsourcing, netting over 60% of the US IT and business process output (BPO) outsourcing market (Sourcingline, 2012). Unfortunately, physical security in India is tenuous at best due to significant terrorist activity related to the ongoing tensions over Indian sovereignty of the Kashmir region. Furthermore, the 2010 Mumbai attacks show that western personnel and assets are also a target when tensions over Kashmir come to a boil. It is also important to note that India is under constant threat from geopolitical tensions between itself and a growing Chinese regional military capability. Data center operations in India would most certainly cease to be available abroad were tensions between the two countries to expand into open warfare.

The second issue that needs to be overcome in outsourcing data to India is the dearth of well-qualified security personnel. A typical salary of only USD $4,000 to $5,000 a year for quality Indian security professionals is not enough to entice a sufficient quantity of high-quality employees that are needed to

Data Centers

16

provide adequate security for sensitive data (Siliconindia.com, 2010). The low pay of security personnel also makes for ample opportunity for low cost criminal bribes and/or corporate espionage in a country where bribes and corruption are a widely-tolerated cultural norm.

Third, evidence is now being found that foreign computer hardware in common use abroad may be intentionally engineered to contain nearly undetectable backdoors that allow corporate and/or government espionage (Skorobogatov & Woods, 2012). Certainly, any data flowing to and from an Indian data center would be at the mercy of your Indian telecom provider’s security efforts. Moreover, depending upon which fiber optic circuits are used to transmit your data it could be transiting mainland China, Pakistan, or even Iran. Of course it could be argued that data encryption provides a solution to this threat, but encryption is often broken with a significant time lapse before the compromise is discovered. Consider for example, that RSA token security was once thought to be a nearly impenetrable security system for secure authentication, but was cracked none the less. Finally, even if data is never compromised, data availability could be disrupted if physical access to offshore data were cut off by an angry neighbor due to geopolitical rivalry.

The fourth issue with security in India is with the appropriateness of Indian employees. While establishing someone’s identity in India is improving due to national biometric identity database (Economist, 2012), vetting Indian employees is still difficult due the accepted practice of bribing government officials as a norm while doing business in India. Ultimately, it is very difficult to know that you have hired the person you think you have hired. Assuming that you contract with a reputable firm that does a good job of vetting Indian employees and has good security practices (such as WiPro or InfoSys for example) you still have a greater risk of insider threat than in a US-based data center. Finally, additional threats exist in India that do not affect US-based employees because kidnapping for ransom, blackmail and other methods of extortion are commonly used in India to leverage an otherwise

Data Centers

17

trustworthy employee and thus gain access to sensitive information. The threat is sufficient that there are even Indian firms that sell insurance discreetly to businesses to offset the financial risk from these threats and compensate loss incurred as a result of sensitive information compromise.

The final and most difficult to overcome security threat with offshore outsourcing is with the legal system of the host country. Companies must know the legal system of country where they are outsourcing their data and a company’s ability to have problems addressed in the host-country legal system. US law simply does not apply and International Law may not apply either. US data security laws requirements such as those found in Sarbanes-Oxley or HIPPA are not binding for an Indian company. Ideas that are stolen are not subject to US Patent enforcement and there is no real guarantee of respect for intellectual property rights outside the United States, Western Europe, and Australia. To overcome this shortcoming many Indian companies seek outside certifications such as ISO 9000 or Carnegie Mellon Software Engineering Institute's Capability Maturity Model (CMM) to reassure potential customers, but companies that adhere to these stringent standards may still have incidents.

Whether an incident is an overt theft or an accident will not matter in the United States. US courts will enforce civil or criminal punishment against the US Company that made the decision to outsource, not the Indian provider. US courts have no jurisdiction in India to recoup damages from intellectual property losses nor will company lawyers have the benefit of precedent when operating in India’s civil law legal system. Finally, the potential loss of confidence in a publically traded company that is perceived to have negligently compromised data security through outsourcing could cause a stock selloff that bankrupts the company. Companies should consider whether the data they are outsourcing is appropriate to be located in a foreign country and protected by the host country legal system.

In conclusion, the above risks represent a significant threat that could negate the financial benefit of an offshore outsourcing decision when realized.

Data Centers Loss of US Jobs:

18

A recent 2012 Sourcing Line survey cited on statisticbrain.com (2012) found that US companies outsourced over 2.2 million jobs to offshore providers in 2011. Of those 2.2 million jobs, 43% were in the IT-sector, with 34% of surveyed US companies stating that one of their reasons for outsourcing was to “gain access to IT resources unavailable internally” and 44% of surveyed companies stating that their motivation was to “reduce or control costs.” Regardless of motivation, the net loss of over 900,000 US IT-sector jobs remains. Finally, it is important to note that it is unknown exactly how many of the 900,000 IT sector jobs were data-center jobs that would be regained if data center services were repatriated.

While the total loss of tax revenue associated with outsourced job loss is difficult to estimate it most certainly is not an insignificant sum. Apple Computer, for example, earned over $24 Billion outside the United States for the Fiscal year 2012 and paid an effective tax rate of only 2.5% on that income (Associated Press, 2012). The continued loss of state, local, and federal tax revenue due to outsourcing translates to degraded essential public sector services that we depend upon as a society--such as police, fire, public infrastructure maintenance, national defense, and public schooling, to name a few. Finally, the loss of tax revenue comes at a time when the US federal budget deficit is reaching critical mass.

Companies that send jobs abroad are beginning to face significant scrutiny from the public and even the US Congress. Recently, progressively-minded lawmakers have even proposed legislation called the Stop Outsourcing and Create American Jobs Act of 2010 (H.R. 5622) with the intent of curtailing the loss of American jobs and tax revenues. Most significantly, the outsourcing activities of Bain Capital may have contributed to the defeat of Governor Mitt Romney, a former Bain Capital executive, in his bid for the 2012 US Presidency. In the current climate, reform to close loopholes that allow companies to realize untaxed offshore profits seems imminent.

Data Centers IP Latency Issues:

19

Analysis of monthly statistics for Latency between North America and India show an average latency of 272 mls with a range from 258 mls to 286 mls (Verizon, 2011). This is reflected in service level agreements for IP transport services to India as well. For example, Verizon will only commit to a latency of 380 mls or less between India and a US destination. Latency this significant could make India inappropriate for many business outsourcing needs. In particular, voice over IP communications are likely to have problems with latency above 100 ms (Bandwith.com, 2012).

Suitability and Benefits of Alaska as a location for a secure green data center:

Location security: Alaska is located at the extreme northwest of North America. The nearest neighbors of consequence to Alaska are Canada and Russia although the portions of Alaska that come close to Russia are non-metropolitan areas of mostly uninhabited tundra and the rocky sparsely populated Aleutian Island chain.

Alaska is home to a large amount of US military capability and is the home of the 3-star Alaska Command (Joint) and 2-star US Army Alaska. Alaska houses 11 military installations with a presence of 15,864 active duty military, 5,556 reservists and 2,766 Coast Guard personnel for a total military presence of around 24,016 (Department of Defense, 2012). A key mission for these forces is to defend mission-essential vulnerable infrastructure including the State’s telecommunications infrastructure.

Overall, there is adequate security and little regional geopolitical threat at present that would threaten a data center located within Alaska. This assumption would need to be revised should a war breakout between the United States and an Asian competitor such as China or North Korea. It is worth noting that, Department of Defense storage requirements for the storage of unclassified data and processing capability represent a potential untapped market for contract data storage. Finally, military needs in

Data Centers

20

Alaska have the potential to grow because global warming is creating a requirement for far-north naval capability to secure the so-called northern shipping route as well as to maintain access to petroleum resources on the ocean floor off the northern coast of Alaska. These requirements have even resulted in initial reconnaissance by DOD officials to evaluate sites for naval and coast guard bases on both the western coast and the northern slope.

Telecommunications Infrastructure: Alaska is connected to the world primarily through four fiber optic cables that traverse the Gulf of Alaska and the Northern Pacific Ocean to land at terminuses on the West Coast of the United States. Additionally, a project is underway to establish a Tokyo, Japan to London, England fiber optic connection through the Bering Strait that would connect Asia and Europe directly. The cable would make landfall at points along the Aleutian chain and Western Alaska where it would connect back to the existing internal fiber-optic infrastructure in the south-central Alaska (Anchorage) region. Proponents of the project claim it would represent a 50% faster data transmission rate for transmission between Asia and Europe. Security for Alaska region undersea cable is provided by the US Navy and US Coast Guard.

Within Alaska there are two commercial fiber optic cables that travel the so-called road belt to connect cities, towns, and military bases located along the road system with broadband capability. One cable travels from the landfall of an undersea cable at Valdez, Alaska and travels north along the highway. The other cable follows the Parks highway from Anchorage. Both cables converge in Fairbanks, Alaska to provide the city of Fairbanks and nearby military bases with dual-homed transmission redundancy.

Capability within the road belt is sufficient that commercial Internet providers such as Global Communications International are able to provide relatively low cost, low-latency (less than 100 ms delay to the East Coast of the US) business and residential broadband (20 MBPS) internet to customers in all of the major Alaskan population centers. Residents in Alaska’s larger cities can easily

Data Centers

21

telecommute, stream movies, use international VOIP telephone service, and make use of any number of other internet technologies with only rare interruptions due to transport problems.

Most connectivity outside of the road belt makes use of DSL over existing telephone circuits that are provided by a combination of high-latency (1000-4000 mls delay) terrestrial microwave relay towers and Satellite connections. Both Terrestrial microwave and satellite are subject to problems when extreme weather such as a blizzard interferes with transmission. North Slope crude oil extraction operations are an example of major business operations that are supported primarily by microwave tower relay and satellite connectivity.

An effort is underway by non-profit Connect Alaska and the State of Alaska to increase broadband penetration in the larger remote towns. Proposals have also been made to the state government to install a fiber optic expansion along the coast called the Northern Fiber Optic Link as part of the Asia to Europe cable that would reach most of the most populated western Alaska towns (Kodiak Kenai Cable Company, 2012). These extension efforts are currently being considered for a combination of state, federal and commercial funding since they could support future crude oil, natural gas, mineral extraction, and planned Department of Defense operations from towns along the western coast. Such a cable would also place Alaska as an ideal location for offshore data centers that could quickly provide service to both Asia and Europe via the new undersea cable.

Overall, the current usage and capacity of the Alaskan fiber optic infrastructure is sufficiently robust to provide services to the lower-48 from data centers located in Alaska. At present, any data center established in Alaska to provide reach back services to the lower-48 would have to be located near the road belt to take advantage of existing internal fiber optic infrastructure. However, the installation of the Northern Fiber Optic Link and the trans-arctic undersea cable will allow all fiber-connected Alaska locations to be considered as potential locations for providing data center services to the US, Asia and

Data Centers Europe. Finally, increased natural gas and mineral extraction activity is a potential Alaska-internal market via existing microwave and an opportunity in the fiber optic construction market.

22

Assessment of Human Capital: Most of Alaska’s 722,190 people live in a few larger cities with approximately 387,894 people living in Anchorage or the nearby Matanuska-Susitna borough area, 97,615 in Fairbanks and 56,369 living on the Kenai Peninsula - primarily in the fishing/tourism economy towns of Kenai, Soldotna, Seward, and Homer (Alaska Department of Labor and Workforce Development, 2012). Of these locations, only Anchorage and Fairbanks have the kind of diversified economic activity that is typical in most metropolitan area US cities.

In general, Alaskans outside of the larger cities are not technically savvy and often embrace a subsistence culture. However, Alaska’s IT talent pool is growing due to a steady influx of US military personnel that stay in the area at the completion of their service obligation. Tech support that moves to Alaska to support crude oil extraction activity is also a source of IT talent. Finally, Alaska has more than ample capability to train needed IT workers for a data-center project within the state of Alaska.

The state is home to three major Universities offering degrees in computer science (two in Anchorage and one in Fairbanks). Additionally, the state of Alaska runs a low-cost resident vocational education program that provides entry-level technicians with CCNA, MCTS, and CompTIA training. Finally, several trade schools exist that provide training for network and server technicians. Combined, these programs create several hundred new IT workers each year. Anecdotally, many of these graduates would prefer to stay in the state, but often move to the lower-48 due to lack of opportunity.

Overall, Alaska does not have a strong base of highly trained IT professionals and the talent it does have is located in the Anchorage and Fairbanks areas. Talent departing from the military also resides primarily in the Anchorage and Fairbanks area. The current downsizing and threat of closure of Eielson

Data Centers Air Force Base near Fairbanks is a potential opportunity to hire high-quality IT workers whose jobs are threatened due to base closure. Romantic notions of Alaskan adventure and frontier spirit have the potential to attract industrious and talented individuals to the metropolitan areas, but probably not in

23

the more remote locations. Finally, potential exists for partnerships with local Universities and the State vocational training program to develop needed data center talent. For these reasons, a data center built in Alaska would have to be located in either the Anchorage or Fairbanks areas to gain access to existing IT talent pools, attract new talent, or develop new talent within the state.

The Impact of Native Corporations: 19.5% of Alaska’s 700,000+ people are at least part Native American or Native Alaskan with the vast majority belonging to the latter (US Census Bureau, 2011). Under the Alaska Native Claims Settlement Act of 1971 most Alaska Natives are represented by 13 regional Native Corporations. The corporations make billions of dollars in land use revenue for the extraction of oil and gas on native lands. Native corporations have also been the beneficiary of over 29 Billion in Federal Funding under 8a Small Business Administration set-aside contracts over the last decade (O’Harrow Jr., 2011). Profits from corporate earnings are used to provide a shareholder dividend payment plan and to support programs beneficial to Native Alaskans, such as universal free health care. In spite of this financial assistance, Native Alaskans remain among the most underserved communities in the United States with respect to education and job opportunities. Any investment that could provide opportunities for Native Alaskan peoples could be eligible for Federal and State Funding that is set aside to promote the development of opportunities for Native Alaskan people. More importantly, the opportunity exists for a partnership with Alaska Native Corporations to leverage access to resources and Native Corporation funding.

Tax Climate: The state imposes no inventory tax, no gross receipts tax, no state sales tax, and no personal income tax. A share of the state’s profit from crude oil extraction fees is paid to state residents

Data Centers each year in the form of the Personal Fund Dividend (PFD). The State of Alaska corporate income tax

24

varies depending on profit with a maximum of 9.4% for profits in excess of $90,000. Additionally, some municipal property taxes exist for businesses in Anchorage and larger cities, but in general, Alaska state and local tax laws are set up to be very attractive to business. Special tax rates for high-earning corporations, such as the petroleum and mining industry, are often negotiated to spur investment. A sizeable tax credit is available for investment/exploration that leads to the development of mineral or petroleum resources.

Climate Impact: Alaska’s area is massive and as such there is a wide variety of temperature fluctuation. With that said, the weather is generally cool due to its northern latitude, with January temperatures regularly dropping to -25 degrees Fahrenheit at night in Anchorage and -45 degrees Fahrenheit in the interior around Fairbanks. Snow typically covers the ground in much of the state from November to April. The warmest month is July, with record temperatures occasionally in the 80s for Anchorage and the 90s in Fairbanks. In spite of the cool climate, personal experience at federal data centers in Alaska has shown that climate control is critical to prevent overheating of IT equipment during summer.

Green Energy Potential: Several projects have demonstrated the Anchorage area’s potential for hydroelectric power and wind power, but unfortunately USGS studies indicate that it is not suitable for geothermal energy production. The Fairbanks area has demonstrated potential for Geo-thermal energy production (SMU, 2004).

Significant Risks:

Environmental threats: The most potentially catastrophic risk in Alaska comes from Earthquakes. In 1964, the Anchorage area suffered a 9.2 magnitude earthquake. The quake leveled the growing town and remains the second largest earthquake on record anywhere in the world (Wikipedia, 2012). Much

Data Centers of South-Central Alaska including Anchorage is also threatened by regular volcanic activity. Ash is a threat to all systems that depend on any kind of air intake such as air-conditioning and internal combustion engines used for emergency power generation. The threat of ash fallout from volcanic activity near Anchorage has forced local military headquarters to move temporarily to the Fairbanks area in the past. It also forced local military leadership to consider shutting down local military data

25

processing centers to prevent damage from ash fallout. Data centers in the Anchorage area would have to address this threat.

Fiber Optic Cable Cuts: Both undersea and underground Fiber optic cable is subject to being cut or damaged by a variety of threats including extreme weather, accidental cuts, or even sabotage/terrorism. Fiber optic cable running between Alaska and the lower-48 would definitely be a target for enemies of the United States in the event of a Pacific theater war.

One example of a fiber optic cable cut disrupting business operations occurred in October of this year when Alaska Airlines had to cancel flights in and out of Alaska after cuts to Sprint’s fiber optic cable in Wisconsin and Oregon caused outages for a key Alaska Airlines system located in the lower-48 (Converge, 2012). Another example occurred in 2006 when a railroad crew accidentally cut the cable running from Anchorage to Fairbanks while operating a few miles north of Talkeetna, Alaska. This occurred at the same time as a cut along Valdez to Fairbanks cable that resulted when heavy flooding caused the nearby road to shift in a narrow mountain pass and sever the cable. The result was a neartotal blackout to commercial communications capabilities in Fairbanks although some limited capability remained through satellite and microwave relay back-ups. The bottom line is that a data center located in Alaska would connect to the world through fiber optic cable that is possibly more vulnerable to cuts that other data center locations because of the limited fiber-optic infrastructure within the state.

Data Centers Chapter 3

26

Proposal and Recommendations: Build a large green-technology 2000-rack commercial data center near Fairbanks, Alaska that will be used to provide competitively priced IT business services and data storage. Leverage the following recommendations to maximize profit and reduce risk:

A. Use the best industry practices for data center thermal design, natural air cooling and the power of stored snow cooling systems to eliminate electricity costs associated with cooling data equipment; endeavor to achieve a PUE of 1:1. B. Establish a partnership with Chena Power to develop and leverage geothermal resources in the Fairbanks area that can provide for 100% of power required by the data center. Build two 10-MW Geothermal Plants that would supply the approximate 8000 kWh of power that would be required on average by the data center and also allow for power source redundancy. Sell excess power generation capability to the Municipality of Fairbanks. Consider development of a natural gas powered backup capability for contingencies that takes advantage of planned State of Alaska strategic plans for development of natural gas fields. C. Work with Doyon, LTD. (the Fairbanks area Regional Native Corporation) to gain low-cost access to land, develop a partnership for outsourcing snow-mound management operations, and to create job opportunities for qualified native technicians and managers. Additionally, once fiber optic cable is extended to the West Coast, consider leveraging native corporations in a secondary business opportunity to establish call centers that would provide beneficial jobs to English speaking Native Alaskans that make up most of the population in west coast Alaska towns. D. Establish a partnership with the University of Alaska to leverage local arctic expertise and establish internships for talented computer science majors. Work with the University to develop an applied research program to study the green data center and optimize operations.

Data Centers E. Work with the Department of Veterans’ Affairs and Alaska National Guard Bureau to develop a

27

hiring program for qualified veterans. Focus on recruiting veterans and Alaska National Guardsmen with technical skills and existing Department of Defense security clearances that are a prerequisite for creating private sector opportunities to store sensitive but unclassified federal and State of Alaska data. F. Leverage talent that will be departing from the Eielson Air Force Base as operations diminish and qualified civilian government workers are displaced. This is another source of employees with existing Department of Defense security clearances that are a prerequisite for creating private sector opportunities to store sensitive but unclassified federal and State of Alaska data. G. Apply for federal grant money and low cost business loans that are available through the Department of Energy for green technology development. Use this to offset start-up costs and generate publicity. H. Develop a product marketing plan that emphasizes the following talking points: a. Environmental sustainability. b. Data Security c. Quality of service provided by native English speaking technicians. d. Providing jobs to Americans including Veterans and Native Alaskan communities. I. Seek additional investors to diversify risk ownership by emphasizing the profit to be made from providing services for international customers via the trans-arctic undersea cable. Financial Analysis:

Construction of a typical commercial data center is estimated to cost approximately $76,000,000 with an annual revenue of approximately $31.9 Million USD per year before maintenance costs at a 75% equipment usage ratio (Seekingalpha.com, 2011). Furthermore, estimated data center maintenance cost is approximated at 3% of revenue per year (Seekingalpha.com, 2011) or $957,000 per year, thus

Data Centers

28

reducing annual return to approximately $31 Million USD. As previously stated, a typical data center is estimated to use approximately 200 million Kilowatt hours of electricity per year (Tsuda, Tano, & Ichido, 2010) at a cost of $12,000,000 USD.

Construction of a geothermal power plant is estimated at $3.9 Million USD per 10 MW geothermal plant ($7.8 Million USD for two) using the previously stated (worst case) cost of approximately $3900 USD per kWh of power generation capability (Alaska Center for Energy and Power, 2012). Annual maintenance costs for a geothermal plant in Alaska are estimated at approximately 2.5% of the plant construction cost for a total of $195,000. The 2.5% estimate is based upon the actual ratio between annual maintenance costs and build costs at the Chena Hot Springs Geothermal power generation plant. The $2 Million facility at Chena costs about $50,000 a year annually to maintain (OpenEI, 2012); approximately 2.5% of build cost. Finally, excess generation has the potential to be sold commercially to the city of Fairbanks.

Expected green data center electricity consumption is expected to be 112 Million Kilowatt Hours with an annual cost of $195,000 USD for geothermal plant maintenance (around $.0017 USD per Kilowatt hour used). The 112 Million kWh figure was arrived at by subtracting the estimated 44% energy usage for airconditioning that is required in a typical lower - 48 data center from the typical 200 Million kWh total usage. The rationale for removing this 44% is that it represents energy that would be saved by using a combination of natural air and snow mound air-conditioning technology. Approximate cost of annual snow field maintenance operations to support the green AC system is estimated to be approximately $1 Million USD (note that this figure is a guess and a separate study should be commissioned to examine the true cost of snow field maintenance).

In summary, total construction costs for a green data center are estimated to be $76,000,000 for the data center, $7,800,000 for two 1 MW geothermal plants, and $1 Million for the snow mound AC system

Data Centers for a total of $83,800,000. Annual operating and maintenance costs are expected to be around $195,000 USD for geothermal maintenance and an estimated $ 1 Million USD for snow field maintenance.

29

Financial comparison between a typical data center and a green data center located in Alaska:

Estimated annual revenue for a geothermal-powered 2000 rack commercial data center with green climate control capability in Alaska is $41,805,000 USD per year. This figure incorporates the $31 Million USD average data center earnings reported by Equinix (Seekingalpha.com, 2011) after maintenance costs plus an annual savings of $10,805,000 USD in electricity costs that would be realized as profit. Electricity cost savings were calculated by subtracting costs for annual geothermal plant maintenance and snow field maintenance operations from the typical $12,000,000 in electricity used by a standard data center located in the lower-48 without geothermal and snow mound AC technology. Overall, a geothermal powered green data center in Alaska has the potential to return $10.8 Million USD more per year to investors than a data center in the lower-48. Finally, potential revenue from the sale of excess energy was not evaluated in this comparison.

Payback Analysis: Payback for a standard data center located in the lower-48 is expected to be realized in 2.45 years with a cost of $76 million USD and annual profits of $31 Million USD, thus generating a 40% annual rate of return on investment. Payback for a green data center located in Alaska is expected to be realized in a little more than 2 years with a construction cost of $83,800,000 USD and annual profit of $41.8 Million USD per year, thus a 49% return on investment.

Chapter 4

Conclusions:

Data Centers Construction of green data centers in Alaska is both economically and technical feasible at this time.

30

The Chena Hot Springs project provides ample evidence to show the feasibility of producing geothermal power in the Fairbanks, Alaska area that will provide for equipment electricity requirements. Case studies at Google and NREL show the feasibility of using Alaska’s naturally cool climate to cool data centers for as much as 9 months out of every year. The Hokkaido data center project in Japan also demonstrates the feasibility of using Alaska’s abundant annual snowfall to leverage snow mound cooling systems that can be used to cool a data center during the few months when ambient temperatures in the Fairbanks area may be too high for natural air cooling. Collectively, the above technologies have the potential of creating a facility that will not only have a PUE of 1:1 but emit no environmentally damaging emissions whatsoever with energy leftover that can be sold to the local community.

Current and planned fiber optic capability within Alaska exists to support adequate connectivity with the continental US with substantially less latency than would be experienced when outsourcing to India, the most popular and likely outsourcing destination. Data centers in Alaska have also been shown to be significantly more secure for US companies when comparing threats to data in Alaska versus to threats in India. The cost of outsourcing to the American economy and services has also been demonstrated.

Local sources for qualified technicians have been tentatively identified with the possibility of establishing beneficial partnerships with several key organizations while providing job opportunities for Veterans and historically disadvantaged Native Alaskans. Substantial benefit to the local Alaskan economy is expected as a result of this project.

Finally, financial analysis shows that the use of green technology to reduce energy requirements results in lower operating costs, faster payback, and a 9% improvement for return on investment. The tax climate has been shown to be one of the most advantageous in the country.

Data Centers References

31

AIDEA - Alaska Industrial Development and Export Authority (2011). Snettisham Hydroelectric Project. Accessed online on 2 November, 2012. Available online at http://www.aidea.org/Programs/ InfrastructureDevelopment/snettishamHydroelectricProject.aspx

Alaska Center for Energy and Power (2012). “Geothermal Reduction in Potential Cost of Energy”. Accessed online on 30 October, 2012. Available online at http://energyalaska.wikidot.com/geothermal-potential-reduction-in-cost-of-energy

Alaska Department of Labor and Workforce Development (2012). “Research and Analysis>Population Estimates”. Accessed online on 18 November, 2012. Available online at; http://www.labor.state.ak.us/research/pop/popest.htm

Associated Press (November 4th, 2012). “Apple Paid only 2 pct on tax earnings outside US”. Accessed on 5 November 2012. Available online at http://news.yahoo.com/apple-paid-only-2-pct-taxearnings-outside-204342044--finance.html

Browntax.com (2012). Curtailing the Outsourcing Epidemic: Lawmakers’ Fight to Bring Back Jobs. Accessed on 6 November, 2012. Available online at: http://www.browntax.com/Legislation/ Curtailing-the-Outsourcing-Epidemic-Lawmakers-Fight-to-Bring-Back-Jobs.shtml

CIRI - Cook Inlet Region, Incorporated (September 2012). “Fire Island Wind Fact Sheet”. Accessed on 5 November, 2012. Available online at: http://www.fireislandwind.com/documents/FIWFactSheet_sep2012_web.pdf

Data Centers Converge Network Digest. (8 November, 2012). “Sprint Suffers Dual Fiber Cut Leading to Outage at Alaska Airlines”. Accessed online on 18 November, 2012. Available online at: http://convergedigest.blogspot.com/2012/10/sprint-suffers-dual-fiber-cut-leading.html

32

Department of Defense (2012). “Alaska Facts and Figures”. Accessed online on 16 November, 2012. Available online at: http://www.defense.gov/specials/outreachpublic/ak.html

Green Grid (2008). . “Green Grid Data Center Power Efficiency Metrics: PUE and DCIE.” Accessed online on 20 October, 2012. Available online at http://www.thegreengrid.org/~/media/WhitePapers/ White_Paper_6_-_PUE_and_DCiE_Eff_Metrics_30_December_2008.pdf?lang=en

Google (2012). “Google Data Center Efficiency: How Others Can Do It - Five Things You Can Do Now.” Accessed on 5 November, 2012. Available online at http://www.google.com/about/ datacenters/efficiency/external/

Holdman, G. (2007). “Fact Sheet: The Chena Geothermal Plant.” Accessed online on 20 October, 2012. Available online at: http://chsr.squarespace.com/storage/documents/Powerfactsheet.pdf

Karmali, L. (October 4th, 2012). “Mists of pandaria pushes warcraft subs over 10 million”. Retrieved on 25 October, 2012. Available online at http://www.ign.com/articles/2012/10/04/mists-ofpandaria-pushes-warcraft-subs-over-10-million

Kodiak Kenai Cable Company (2012). “Northern Fiber Optic Link”. Accessed online on 20 October, 2012. Available online at http://www.northernfiberlink.info/

Manyika, J & Roxburgh, C. (October, 2011). “The Great Transformer: The impact of the Internet on economic growth and prosperity.” Accessed on 20 October, 2012. Available online at:

Data Centers

33

http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&cad=rja&ved=0CCgQFj AB&url=http%3A%2F%2Fwww.mckinsey.com%2F~%2Fmedia%2FMcKinsey%2Fdotcom%2FInsig hts%2520and%2520pubs%2FMGI%2FResearch%2FTechnology%2520and%2520Innovation%2FT he%2520great%2520transformer%2FMGI_Impact_of_Internet_on_economic_growth.ashx&ei= EuKKUM7AKoulqQHh6oHACA&usg=AFQjCNFlbvJ7_U8QND5PMmGMGcCYhnxf-A

McCarthy, J. (2003) Unlocking the Savings in Offshore. Forrester Research 2003.

National Renewable Energy Laboratory (December, 2011.) Reducing Data Center Loads for a Largescale, Low-energy Office Building: NREL’s Research Support Facility. Accessed on 12 October, 2012. Available online at: http://www.nrel.gov/sustainable_nrel/pdfs/52785.pdf

O’Harrow Jr. , R. (February 11, 2011) Small Business Administration changes rules for set-aside contracting program. Washington Post. Accessed online on 11 November, 2012. Available online at http://www.washingtonpost.com/wp-dyn/content /article/2011/02 11 /AR2011021106212.html

OpenEI (2012). Resource Area. Chena Geothermal Area. Accessed on 23 November, 2012. Available online at http://en.openei.org/wiki/Chena_Geothermal_Area

Sandvine Intelligent Broadband Networks (Fall, 2011). “Global Internet Phenomenon Report”. Accessed on 23 October, 2012. Available online at http://www.sandvine.com/downloads/documents/1026-2011_phenomena/Sandvine%20Global%20Internet%20Phenomena%20Report%20%20Fall%202011.PDF

Scott, D. (2010). “Kodiak Electric Association Pillar Mountain Wind Project.” Accessed on 20 October, 2012. Available Online at: http://www.edinenergy.org/pdfs/ce_workshop2010_scott.pdf

Data Centers

34

Seekingalpha.com (April, 28th 2011). “Equinix's CEO Discusses Q1 2011 Results - Earnings Call Transcript” Viewed online on 18 November, 2012. Available online at http://seekingalpha.com/article/ 266082-equinix-s-ceo-discusses-q1-2011-results-earnings-call-transcript.

Shelly G. & Rosenblatt, H. (2012) Systems Analysis and Design, NINTH (9TH) Edition.

Skogsberg K. and Nordell B. (March, 2001). “The Sundsvall hospital snow storage”. Cold Regions Science and Technology. Volume 32; Issue 1, pp 63-70.

Skorobogatov, S. & Woods, C. (March, 2012). Breakthrough Silicon Scanning Discovers backdoor in Military Chip (Draft of 05 March, 2012). Accessed online on 6 November, 2012. Available at: http://www.scribd.com/doc/95282643/Backdoors-Embedded-in-DoD-Microchips-From-China.

Silicon India (September 15th, 2010). “Indian IT's physical security at stake: No manpower.” Accessed online on 6 November, 2012. Available online at http://www.siliconindia.com/shownews/ Indian_ITs_physical_security_at_stake__No_manpowem_-nid-71633-cid-2.html.

Sourcing Line (2012). “India”. Accessed online on 6 November, 2012. Available online at http://www.sourcingline.com/outsourcing-location/india.

SMU - Southern Methodist University (2004). “SMU Geothermal Map of Alaska”. Accessed on 1 November, 2012. Available online at http://smu.edu/geothermal/heatflow/Alaska_hf.gif

Stansberry M. and Kudritzki, J. (2012). Uptime Institute 2012 Data Center Industry Survey. Accessed on 1 November, 2012. Available online at: http://uptimeinstitute.com/images/stories/ Uptime_Institute_2012_Data_Industry_Survey.pdf

Data Centers Statisticbrain.com (20, July, 2012). “Job Outsourcing Statistics”. Accessed on 6 November, 2012. Available online at: http://www.statisticbrain.com/outsourcing-statistics-by-country/

35

Tsuda K., Tano S., & Ichido, J. (2010). “Lower Data Center Power Consumption through Use of the Climate Characteristics of Cold Regions and Inter-regional Energy Integration.” Proceedings of the 2010 IEEE International Conference on Progress in Informatics and Computing PIC 2010. Volume 2.

US Energy Information Administration (September, 2012). “Electrical Power Monthly with Data for July 2012.” Accessed on 24 October, 2012. Available online at: http://www.eia.gov/electricity/ monthly/pdf/epm.pdf.

US Census Bureau (2011). American and Indian Native Heritage Month 2011: Facts and Figures. Accessed on 10 November, 2012. Available online at http://www.census.gov/newsroom/ releases/pdf/cb11ff-22_aian.pdf

Verizon (2011). “IP Latency Statistics”. Accessed online on 23 November, 2012. Available online at http://www.verizonbusiness.com/about/network/latency/

Wikipedia (2012). 1964 Alaska Earthquake. Accessed online on 18 November, 2012. Available online at http://en.wikipedia.org/wiki/1964_Alaska_earthquake.

Data Centers Appendix A: System Request Sheet

36

Data Centers Appendix B: Project Definition Sheet

37

Company: Big Oil International Corp. Division: Project Management Office Project Title: Sustainable Green Data-Center

Project Manager: Jason Murray Sponsor: Chief Financial Officer Project Code: ITM6000Murray12

Problem Statements: 1. Data Center Access: Our company currently requires significant secure data storage for our

Alaska based operations that is expected to increase as we begin to extract natural gas in addition to our current crude operations. Current and projected remote work sites are completely dependent upon reach-back to data-centers located in the lower-48 states that provide access to critical application server services and the company ERP system. The use of low-cost Asia-based data centers is not desirable because of concerns about company proprietary data security and access to Asia based datacenters requires transport back to the lower-48 regardless. As a result the decision has been made to consider the feasibility of a company data center within the state of Alaska

2.

Data Transport: Current reach-back to lower-48 requires significant transport outside the state

across limited bandwidth leased terrestrial circuits that ride one of four undersea fiber optic circuits to sites in Washington State and Oregon. The current fiber optic circuits available for lease are insufficient for the pace and volume of current petroleum and projected natural gas operations. The possibility of using satellite based technology to provide transport in remote areas is too limited and has become too costly. Most of our current Satellite connectivity comes from leased channels on a commercial satellite that is expected to degrade from its current orbit in approximately 5-years at which time we will have very limited options with respect to space-based transmission. A new satellite positioned appropriately

Data Centers for our required far north look angle is not expected to be launched and is not considered to be commercially viable. The only options for long-term internal transport and adequate external bandwidth are microwave and fiber-optics. Microwave is used extensively, but remote microwave

38

antenna sites must be powered and maintained at great cost and are often degraded completely during white out conditions during the winter months. With great consideration to long term operations and sustainable costs the decision has been made to evaluate the feasibility of investing in undersea and terrestrial fiber optic enhancements within the State of Alaska to support company data transport infrastructure requirements. 3. Additional Strategic Goals: A. The company would like to avoid costly data center energy costs by building a completely energy self-sufficient data center using cutting edge green energy technology. B. The company would like to offset the costs of building the data center with federal research funds and tax incentives set aside for green energy. C. The company would like to generate carbon tax credits that can be used to offset other company operations or sold at a profit to less efficient companies. D. The company would like to realize public goodwill for investing in green energy and providing green tech-sector jobs. E. The company would like to use existing State of Alaska funded programs for training technicians to prepare a workforce capable of operating green data centers and maintaining fiber optic circuits. F. The company would like to realize incentives for employing and leverage GI Bill funds in developing workers with at the same time taking advantage of workers with existing federal security clearances. G. The company would like to develop and patent new proprietary green energy technology that will ensure long-term profits and energy sector dominance as petroleum resources decrease.

Data Centers

39

H. The company would like to leverage funds for State of Alaska initiatives to expand broadband access into remote areas that can be used to offset costs for needed fiber optic cable infrastructure enhancement. I. The company would like access the significant land and financial resources of Alaska Native Corporations in building the new data center and realize public goodwill for providing quality job opportunities to a historically disadvantaged minority community. J. The company would like to partner with local Universities to develop new technologies, leverage arctic research expertise/funding, and develop the IT workforce. K. The company would like to be able to leverage the data center to provide outsourced secure datacenter services to public and private sector that will not only offset operational data-center costs but generate a profit. L. The company would like to profit from the lease of circuits for fiber-optic cable infrastructure that installed. Project Goals: Mandatory: Provide a feasibility study that examines the leveraging of existing green technology to build a green data center in Alaska and provide recommendations for fiber optic cable expansion that would support such a data center. Desirable: Provide examples and estimates of potential profits from additional business opportunities that can be realized from investing in green data centers and expanding fiber optic capability. Priority: Medium. Budget: This is an initial feasibility research study for senior management using publically available data and existing research. It should be approximately 20-30 pages in length and address the problem statements and strategic goals listed above. No more than $100 will be used to purchase necessary

Data Centers

40

research references (studies, journal articles, etc). Funds for additional research or a full-fledged project will be provided at a later date dependent upon senior management’s assessment of the initial feasibility study. Expected Time/Business Deadline: The feasibility study is required for delivery on 7 December, 2012.

Data Centers

41

Appendix C: Abstract This paper reviewed secondary research and case studies to examine the economic and technical feasibility of building data centers in Alaska by demonstrating how the combined use of best practices, green technology, and the unique collection of resources found in Alaska can be leveraged to improve profit and lessen environmental impact. The paper also used India as a basis of comparison to demonstrate how Alaskan offshore outsourcing offers an improved quality of service, better security, and a more advantageous position in the current domestic political climate than does foreign offshore outsourcing.

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close