Windows Server 2003 R2 Enterprise Edition: A
Strategic Choice for an Optimized Infrastructure

Published: January 2007

For the latest information, please see http://www.microsoft.com/windowsserver2003/enterprise

Contents

1

Introduction 1

Objectives of Server Consolidation 2

Approaches to Server Consolidation 2

Target Workloads 8

Benefits of Using Windows Server for Resource Optimization 10

Summary 12

Related Links 13

Introduction

Network infrastructures in many organizations exhibit complex, widely distributed topologies that are the result of decades of evolution. This phenomenon, often referred to as server sprawl, carries with it many hidden costs. These indirect costs are manifest as cash expenditures by the organization and as management time and effort. Over a period of many years, this continual but multidimensional evolution can result in corporate networks with servers sprawling within and beyond the walls of the data center, draining resources and adding cost in a way that is totally contrary to the purpose of the technology. Still, there is ever-increasing demand for server resources as computer applications become a universal component of business processes. Even workers with minimal technical knowledge know that server applications provide a convenient way to deliver information broadly and capture valuable data.

Business justification for servers usually does not take server efficiency into consideration. Generally, a team can justify the cost of a server if it can show that the new server will provide desired benefits such as rapid return on investment (ROI), better collaboration, enhanced productivity, or improved communications with customers or partners. However, a large number of inefficient servers can have massive cost implications beyond the cost of the hardware and software. Servers have other indirect, fixed costs such as power, cooling, rack and floor space, and the time and effort required for management and maintenance. Thus, server efficiency is a product of hardware utilization, administrative effort, and complexity of solution support.

Many servers are deployed with the expectation that they will have very low utilization—as little as 15% in some cases. Regardless of utilization, all servers use costly, valuable datacenter space and add to power and cooling requirements. A large quantity of servers requires a large administrative effort to maintain the hardware and software. All server operating systems need to be regularly updated and require management such as backup and restore procedures. In addition to internal administrative staff, you have an ecosystem of hardware vendors, software vendors, and IT services specialists. These external vendors and service teams require additional time and amplify the support challenge. The cumulative cost of these inefficiencies can be staggering. Over the life of a server these indirect costs can be as high as 6 times the original cost of the server hardware and software.

It is therefore in the interest of any business to consider consolidating its servers as a means of eliminating wasted resources and containing server sprawl. Eliminating low utilization servers by consolidating workloads onto fewer servers can result in massive cost savings due to reduced datacenter demands, reduced administrative effort, and simplified solution support. Business units experience no negative impact on results when their workloads are hosted on servers that are running at higher utilization, and the organization gains the added benefit of greater IT efficiency. Thus, it is not surprising that many businesses are already undertaking server consolidation initiatives and seeking alternatives to the conventional, often undisciplined approach to server deployment.

Microsoft provides a solution that incorporates both software and prescriptive guidance to deliver a platform and a plan for server consolidation that helps to eliminate low server utilization and to control server sprawl. This can lead to a more efficient use of hardware, more productive and responsive IT teams, and simplified support. The Microsoft® Windows Server® 2003 R2 Enterprise Edition operating system can help you lower the cost and complexity of managing your server infrastructure.

In this paper, we assess the problems of server sprawl and underutilization and describe how deploying Virtual Server 2005 R2 with Windows Server 2003 R2 Enterprise Edition provides your organization with a secure, reliable platform that can help you meet and manage demand for server resources more efficiently.

Objectives of Server Consolidation

The fundamental objective of any server consolidation initiative is to enable the business to save money by improving hardware efficiency, increasing the productivity of IT staff, and simplifying support.

Direct cost savings can be realized by minimizing server sprawl, which reduces hardware cost arising from the deployment of new, under-utilized servers. By remodeling organizational approaches to IT with a focus on reducing server count and increasing server utilization, enterprises can realize superior value per server, lower total cost of ownership (TCO), and faster return on investment (ROI). Reducing server count also helps alleviate the indirect cost of data center real estate and its associated support burden.

In today’s IT environment, technology budgets and staff are being reduced, yet SLA (service Level Agreement) requirements are rising. Often, the only way you can meet these seemingly contrary demands is to consolidate and optimize your infrastructure. A streamlined server support and management infrastructure minimizes cost by alleviating IT resource demands. An IT team is most valuable when it focuses its efforts and expertise on strategic initiatives, rather than struggling to manage a sprawling, disorganized server infrastructure. Windows Server 2003 R2 Enterprise Edition provides many tools to assist IT teams who want to reap the benefits of improved IT efficiency.

Generally speaking, server consolidation is a strategic, forward-thinking initiative. For most organizations, it is a delicate undertaking that requires careful planning that thoroughly assesses the current infrastructure in the context of projected future needs. Hardware and software standardization enables businesses to further alleviate IT resource strain by optimizing their vendor ecosystem. This requires an organizational approach to new server deployment and a plan for dealing with existing server hardware and software assets.

Businesses can save time and money if they select a software platform that facilitates effective server consolidation. Windows Server 2003 R2 Enterprise Edition offers a standardized platform that is reliable, scalable, and appropriate for universal deployment.

Approaches to Server Consolidation

Different server workloads utilize computing resources in completely different ways, and therefore require different approaches to server consolidation. For example, it would not be appropriate to host a public Web site on a computer that also hosts a business-critical database. This is an example of heterogeneous workload consolidation, which can be suitable for some, but certainly not all, circumstances.

Unfortunately, it is not always as easy to recognize incongruous server workloads as in the example of Web servers and database servers. Even with good prescriptive guidance, you should carefully consider your organization’s operational environment, growth expectations, industry standards, and regulatory requirements as you plan and execute your server consolidation strategy.

While it is common to undertake server consolidation for the purposes of eliminating underutilized server hardware, the objective is really to optimize servers. In this way, a server consolidation initiative can also play a key role in your strategy for new server deployments. Ideally, a server consolidation strategy accomplishes both elimination of low utilization servers and optimization of server resources ongoing.

Common strategies for server consolidation include the following:

/ ·  Centralization involves relocating servers to fewer sites, reducing the number of locations where servers reside. Centralization is usually among the first steps taken by organizations when implementing server consolidation initiatives.
/ ·  Physical consolidation involves replacing many underutilized or small-capacity servers with fewer, more powerful computers or clustered systems.
/ ·  Data integration entails combining data from different sources into a standard format or repository. When implemented properly, this type of consolidation centralizes management, reduces redundancy, and can yield better control of data.
/ ·  Application integration involves combining multiple applications onto fewer servers. This type of consolidation can decrease complexity and redundancy by reducing the number of architectures and platforms that IT staff must support.

To accomplish all of these goals, Microsoft recommends three general approaches to help guide the development of your server consolidation initiatives:

·  Hardware partitioning

·  Workload consolidation

·  Virtualization

These three approaches can help you make the most of your hardware and software investments while maintaining sensitivity to the needs and nature of the workloads being considered for consolidation.

Hardware Partitioning

The first approach to server consolidation is hardware partitioning, in which the physical computer can be partitioned into multiple logical servers that are all running on the same physical server computer. Hardware partitioning requires specialized server computers designed for this purpose. You can see an example of server consolidation using hardware partitioning in Figure 1.

Figure 1. Server consolidation using hardware partitioning.

Hardware partitioning enables you to logically divide a computer into multiple partitions, where each partition is a unique instance of a complete operating system running as an isolated computer. Each partition is allocated computing resources such as processors, memory, and disk storage space. The number of partitions that can be supported depends on the computer hardware and the computing resources available in the physical server. Some high-performance computer hardware can supports as many as eight or more partitions.

Because partitions run separately and independently, you can run different, unique operating systems on each one. For example, one partition can run Microsoft® Windows Server® 2003 while another partition could be running a different operating system such as Sun Solaris or SuSE Linux.

This solution has the additional capability of delivering high-availability configurations within a single server—sometimes called a “cluster-in-a-box.” In this configuration, each partition operates as a separate logical system, giving you the ability to set up a Microsoft Clustering Services node on each of the different partitions in the server computer, thus enabling failover between different partitions. The feasibility of this approach is somewhat limited by your specific server hardware configuration. (You can find comprehensive information about server hardware configurations that are compatible with Windows Server 2003 R2 Enterprise Edition by visiting the Windows Server Catalog of Tested Products Web site at http://www.windowsservercatalog.com/).

Note: Hardware partitioning requires specialized computer hardware which can be expensive. The maximum number of partitions supported by the hardware is another limitation to consider.

Because many of these specialized servers are built on highly scalable 64-bit processor architecture, they can take advantage of server operating systems that are optimized for 64-bit platforms, such as Microsoft® Windows Server® 2003 R2 64-bit Enterprise Edition. Additionally, Windows Server 2003 R2 Enterprise Edition includes features and capabilities for high-availability scenarios as described earlier in this document in the section titled Increased Administrative Effort.

Hardware partitioning is ideal for heavy workloads and for workloads that require full access to computing resources in a totally isolated environment. Examples of these types of workloads include database servers, specialty enterprise business applications, and e-mail servers. Windows Server 2003 R2 Enterprise Edition provides the features and the scalability to ensure these workloads can maximize utilization of computers capable of hardware partitioning.

Workload Consolidation

Another approach to server consolidation is through workload consolidation, also commonly called rational consolidation. In workload consolidation, you run multiple, heterogeneous applications simultaneously on a single instance of the operating system. Because it is very effective in quickly containing server sprawl, workload consolidation is a widely used and compelling approach to server consolidation. However, in some circumstances, workload consolidation can be the most complex way to implement server consolidation.

The workload consolidation approach requires an organization to assess its available server platforms, considering important factors like server location, capacity, configuration, and capability. The IT staff must also think about the types of workloads and, perhaps more importantly, the requirements and behavior of the workloads that are running on the various server computers. In large and complex organizations, this can be extremely difficult to map accurately. You can see an overview of workload consolidation in Figure 2.

Figure 2. Example of server consolidation through workload consolidation.

Once you have an accurate assessment of the server resources, workloads, and requirements involved, you can begin the task of determining the best way to combine the workloads with available server resources while still meeting business needs. For this reason, workload consolidation can be an especially effective, beneficial approach in scenarios where servers have proliferated in a haphazard fashion. For example, workload consolidation is ideal for a department or branch office where there might be multiple single-purpose servers with low utilization that can be consolidated.

The level of effort required to complete the evaluation and planning associated with workload consolidation should not be underestimated, and not all workloads can or should be combined onto shared physical computers. Workload consolidation is best for applications with medium to high throughput that can coexist with other applications. Due to application conflicts that can arise from coexistence, some software manufacturers recommend against this approach.

Workload consolidation can also have political ramifications as different organizations within your company may be hesitant to relinquish the control they have over the computer hardware and software acquired with their budget.

Windows Server 2003 R2 Enterprise Edition helps to overcome many of these challenges. Advances in the operating system make it easier for servers to run multiple concurrent applications. Plus Windows Server 2003 R2 Enterprise Edition includes new, innovative tools that are specifically designed to help manage consolidated workloads.

Microsoft® Windows Server® 2003 Datacenter Edition provides a particularly scalable way to accomplish workload consolidation. Certain kinds of software, including products like Microsoft® SQL Server™ 2005 or Microsoft® Exchange Server, can put a significant load on a server that can demand all available computing resources. With workloads like these, it can be more beneficial to consolidate the applications and data onto a single large server.