Examining the Role of the Data Center

What happens to the data center when companies begin to implement hybrid clouds? First, the data center does not go away. After all, almost all medium-size and large companies run their own data center — which is how many companies operate their systems of record, including accounting systems, payroll, human resources applications, and line of business applications, to name a few. Many data centers have grown in an unplanned manner over many decades. The typical data center supports different hardware architectures, operating systems, applications, and hundreds, if not thousands, of different tools. To make matters worse, a lot of the money spent in supporting a data center is used for maintenance of existing systems, heating, air conditioning, floor space, and labor.

So, it’s not surprising that many companies have taken the time to streamline their data centers through technologies such as server virtualization. In essence, virtualization decouples the software from the hardware. In decoupling, the software is put into a separate container so that it’s isolated from the underlying operating system. (See Chapter 16 for more details on how virtualization works.) With the use of virtualization, data center management can more easily and efficiently manage the way applications are placed on servers.

However, even though IT has made the data center more efficient, cloud computing has made it apparent that more can be done to transform computing. Organizations are beginning to take a hard look at what the centralized data center is well suited for and at the changes required to create a computing environment that truly serves the needs of their business.

Companies are beginning to discover that they can have the best of all worlds by finding the tasks that are most appropriate for a highly controlled data center and which workloads are best suited to either a private or a public cloud.

Although a well-designed and well-tuned data center provides essential services to a company, it is often best suited for a complex line of business applications. These are often transaction-intensive applications that need to confirm and track the movement of financial transactions among customers, suppliers, and partners. Additionally, large, often highly customized systems of record are and will continue to be data center–based. These applications are typically tightly monitored for corporate governance and compliance.

The key difference between the traditional role of the data center and the new role of the data center is just beginning to be clear: The traditional data center is changing from a general-purpose repository for all applications to a highly tuned corporate asset for a class of applications. This is explained in Tables 1-1 and 1-2.

Table 1-1 Traditional Data Center

Type of Data Center	Description
Traditional Data Center	Data center with all applications and data centrally managed
Consolidated Data Center	Virtualization applied to enable server consolidation supporting existing model

Table 1-2 Next Generation Data Center

Type of Data Center	Description
Purpose-Built Data Center	Highly tuned data center for enterprise systems of record
Purpose-Built Private Cloud	Self-service resource available for developers and partners

Traditional data center

One way to think about where we are today with data centers is to think about the car garage in the typical suburban home. When the home is first built, the homeowners use the garage for its intended use — storing two cars. For the first year or so everything works well, but over time the garage is used to store the lawn mower, boxes of books, and random things that are no longer used. Soon there is little room in the garage for even a single car. Figure 1-2 illustrates this concept. The built for purpose garage is no longer efficient and no longer serves its purpose, but the family still needs to store cars and still needs room for everything else. The solution could be to build a shed in the backyard to store yard tools and other stuff. What is the point of this example? It mirrors what has happened to the data center over time.

For the past 30 years, organizations have had traditional data that has been in a centrally managed environment for a huge variety of applications and company requirements. Consequently, the typical data center has become more complicated, expensive, and cumbersome to manage. This situation has resulted in data center consolidation, where IT management has taken a hard look first at what applications are actually needed.

Rethinking the data center

IT management has begun to pare down the number of applications being used, and to use virtualization to consolidate workloads and remove superfluous hardware. Traditionally, it was complicated to determine how much physical space an application would require in the data center. Rather than take the risk that an application would not have enough room to support customer requirements, companies simply bought additional capacity in order to ensure performance. However, this approach to applications management became unsustainable. In essence, through virtualization technology, IT has been able to add a software layer that allows applications to be more easily consolidated onto specific servers. But even these moves have not been enough.

The purpose-built data center

Just as the homeowners reclaimed the garage, IT is beginning to make the transition to purpose-built data centers. These new generation data centers are tuned to support the key systems that manage a business. In some situations, key applications are simply maintained in their current state because of the way they’re used for stable, mission critical business processes. In other cases, the IT organization has taken the time to rearchitect these applications into a set of modular services that can be used to support many different business initiatives. For example, a company might create a service as the authorized way to pay a supplier for a service. By creating a single service that is used across many business units, the data center becomes a more efficient environment. In addition, the emerging purpose-built data center doesn’t try to provide all services for all needs. It becomes the source of the business services and data services that apply to the cross-organizational needs.

Purpose-built private cloud

IT organizations have discovered that it’s much more efficient and effective to create private cloud services for developers to create new applications and services. Therefore, companies are setting up a highly automated computing environment enabled with a self-service portal. This portal is often designed with business process rules that dictate what services a developer or an authorized partner can use. For example, a developer beginning to develop a new application may be permitted to use the Java language, specific types of middleware, and a specified amount of computing capacity and storage. Once the project is completed, there may be a rule that automatically returns the capacity back to the pool of resources. The private cloud service is intended to support an organization’s need for speed and agility based on fast-changing company requirements. Having a private cloud available for projects allows a company to easily experiment with new ideas and new applications without having to request funding for a project that might not become a reality.

Seeing how the public cloud fits

Say that your company is a retailer with a vibrant and well-used transactional portal for selling products online. The biggest event for the company is its yearly 50 percent–off sale. During a two-week period, the company has an agreement with a public cloud company to supplement its capacity so that performance is always consistent. Your company has a contract with the public cloud provider to provide capacity services during the sale as well as other times when additional resources are required. Because your company can rely on the public cloud provider, it doesn’t have to buy additional servers for incremental usage, thus saving time and money.

Your company has selected two different software applications from service providers to replace existing on-premises applications. One application is a customer relationship management (CRM) that allows the sales force to easily get access to prospective and current customer information from a cloud-based public service. In addition, your company uses a human resources management platform as a service. The company has implemented integration software that allows data to be managed among the data center and the two SaaS environments. Both of these SaaS applications have enabled your company to avoid purchasing additional hardware and software that would require IT management. In addition, because the sales team can access their data much faster from any device they’re using, from a business perspective, the sales team’s performance is much more effective.

Your company is in a new market where getting a series of new services operational quickly has the potential to leapfrog the competition. Your company hasn’t been in business very long. Because the company can use a public service, it can make services available before more established companies can act. At the same time, the company can use its data center to monitor the effectiveness of services, manage data privately, and combine with other services that aren’t visible to customers.

Knowing when the private cloud shines

Now, say that your company’s products and services are offered in a portal to your business partners. These products are key to a company’s revenue. For example, you might be part of a financial services company that offers a key business service that is purchased by various banks around the globe. Because this service is a revenue source and because the company is a well-run and orchestrated computing environment, it makes sense for the company to establish a private cloud to support this business model. An architected private cloud that supports this business initiative is ideal for the financial services company. Using a public service where the company is charged on a time or usage basis wouldn’t be cost-effective.

Your company has a large team of developers spread across the world. You decide to create a private cloud that allows any authorized developer to gain access to data, tools, and processes required to create applications. You have set up rules to ensure that developers can access only the tools and data they’re supposed to use. The private cloud makes the process of managing a distributed development team more productive and more cost-effective.