Google and Their Data Center Strategy

Connect the Dots for Your Company

Network and Hardware Section of the Google Strategic Plan

The following network and hardware section of the Google strategic plan is predicated on how the Google business model will increasingly concentrate on cloud computing as a core sector of organic new growth. Reducing the overall cost of ownership, increasing reliability and service responsiveness, and increasing the capacity to quickly launch new services is critical to attaining long-term revenue and profit objectives (Google Investor Relations, 2014). Creating a solid technological foundation for Google to continue expanding on globally is also a strategic objective of this plan. Inherent in any design of Google strategic information technology (IT) plan is contingency planning and back-up plans to ensure systems availability and long-term reliability (Google Investor Relations, 2014). This section of the Google strategic plan also includes key components of this aspect of strategic IT planning as well.

Introduction

Orchestrating the global network of hardware, systems and services to ensure continual stability, reliability and fault tolerance so business strategies can be supported and sustained is the goal of the Google IT Strategic Plan. The network and hardware section's goal is to concentrate on how all available IT resources can be used for supporting and strengthening the strategic plans, initiatives and strategies of the company (Google Investor Relations, 2014).

With cloud computing becoming an increasingly larger percentage of Google's annual revenues and profits, this section of the Google strategic plan concentrates on optimizing the network and hardware infrastructure to optimize AppEngine performance. The Google AppEngine has the potential to dominate enterprise cloud computing given its inherent scalability, support across the Google series of fabric controllers that optimize searches, and deep Application Programmer Interface expertise (Prodan, Sperk, Ostermann, 2012). One of the most critically important strategic goals of this plan is creating an optimal architecture to streamline and ensure the highest performance possible for Google AppEngine running in heterogeneous, complex enterprise software environments. This requirement is predicated on the depth and breadth of applications that many Google enterprise-class customers have and the ubiquity of support necessary for them to be productive building AppEngine-based applications. The network and hardware infrastructure needs to reliably and securely scale to support Google AppEngine as a result.

A second strategic initiative that the Google IT plan must support and continually scale to optimize is the corporate-wide performance of the Platform-as-a-Service (PaaS) services, which have grown into a significant business over the last two years (Google Investor Relations, 2014). PaaS Services provide the potential for Google to move beyond its heavy dependence primarily on advertising and shift the focus to competing directly with Amazon Web Services, IBM and others who are defining the broader enterprise cloud computing landscape today (Beimborn, Miletzki, Wenzel, 2011). Creating an optimized IT infrastructure to enable excellent Google AppEngine performance in enterprises (Prodan, Sperk, Ostermann, 2012) and also creating one that can support scalability and enterprise security requirements for PaaS-based platform development and sales is essential for Google's long-term profitability and financial growth (Google Investor Relations, 2014).

Google Network and Hardware Planning: A Global Perspective

Google's global business model needs to be the foundation that the IT strategic plan is built on, starting with the data center strategy. Defining the optimal data center strategy will in turn define the optimal level of network services, supp0ort, servers and application support. As the Google AppEngine and PaaS-based businesses continue to accelerate, it is anticipated that additional data centers will be required. Figure 1 provides an overview of the twelve data centers is existence today.

Figure 1: Current Data Center Locations as of September, 2014

Source: http://www.google.com/about/datacenters/inside/locations/index.html

The first Google data center was constructed in 2003, in Douglas County, GA, followed by locations in Oregon, North and South Carolina and in 2008, the first European data center in St. Ghislain. Belgium (Google Investor Relations, 2014). Currently Google has approximately 2.4M servers installed across 12 data centers worldwide. Table 1, 2014 Server Count by Data Center provides an overview of server count by location.

Table 1: 2014 Server Count by Data Center Location

Source: Analysis of Securities and Exchange documents filed by Google and Investor Relations, 2014)

Google's cloud business is growing at a consistent 20% a year, with the majority of cloud computing enterprises being located throughout the United States. With the majority of cloud computing growth being in North America, continued investment in commdot8y class servers that include virtualization options for scaling over excess, unused computing capacity is a requirement for driving greater Return on Invested Capital (ROIC) relative to strategic planning assumptions in the Google operating budget (Google Investor Relations, 2014). With the 20% assumption of server and hardware growth, Google senior management needs to budget for an incremental 1,743,360 servers over the next five years to support the growth of cloud-based business models and new services. Table 2, Projected Server Requirements for Incremental Growth for Cloud Computing Expansion provides a forecast of IT system requirements overt the next five years to support the revenue and profit objectives for the cloud computing business.

Table 2:

Projected Server Requirements for Incremental Growth for Cloud Computing Expansion

Source: Analysis of Securities and Exchange documents filed by Google and Investor Relations, 2014)

Within five years all of these systems will need to be fully compliant with internal virtualization standards and requirements while also having the ability to be fully complaint to Google Mesh architectural standards. The continual efforts of Google Engineering to define virtualization standards and mesh computing algorithms shows potential to drastically reduce ROIC in these servers over a seven-year depreciation period as well (Google Investor Relations, 2014). Google Mesh standards and APIs will also support ongoing efforts to expand into enterprise cloud computing in companies dominated by a wide variety of diverse applications and systems that need to also be cloud-enabled (Paul, Jain, Samaka, Pan, 2014).. This approach to creating a unified cloud computing architecture will require intensive investments in servers and predominantly open source operating systems which Google has historically standardized on to save on licensing costs (Google Investor Relations, 2014). The Google File System and Grid, which is the foundation of the search engine, apps platform, Google ID membership system, APIs and application layer of the architecture must all be orchestrated together to ensure scalability and security over time. This strategic plan concentrates on how to forecast server requirements across all current and future data centers as the cloud computing business is projected to grow significantly over the next five years.