Research Publication Date: 4 August 2006 ID Number: G00142079 Get Ready for the Revival of Large Centers Rakesh Kumar Although large data centers declined in importance in recent years, they will return to prominence during the next decade, as IT service delivery moves to a more-centralized model. However, the topology of future "mega data centers" will differ from those of the past. Reproduction and distribution of this publication in any form without prior written permission is forbidden. The information contained herein has been obtained from sources believed to be reliable. Gartner disclaims all warranties as to the accuracy, completeness or adequacy of such information. Although Gartner's research may discuss legal issues related to the information technology business, Gartner does not provide legal advice or services and its research should not be construed or used as such. Gartner shall have no liability for errors, omissions or inadequacies in the information contained herein or for interpretations thereof. The opinions expressed herein are subject to change without notice.
STRATEGIC PLANNING ASSUMPTION(S) The number and influence of mega data centers for enterprise customers will increase during the next 10 years (0.7 probability). The requirements for flexible infrastructures (such as servers and storage) and facilities, including power and air conditioning growth, will mean that many of the data center sites developed during the past 20 years ago may be insufficient for the next 10 years (0.8 probability). Technological improvements and streamlined operational processes will result in a physically integrated and logically centralized data center topology during the next five to seven years (0.7 probability). ANALYSIS Server rationalization, hardware growth and cost containment trends are driving the consolidation of enterprise data-processing sites into larger data centers. The number and influence of these large data-processing centers, or "mega data centers," will increase during the next 10 years (0.7 probability). The topology of these sites will differ from the large data centers of the past: A number of factors, including new requirements for flexible infrastructure (such as servers and storage) and facilities (such as power and cooling) and the difficulty of shutting down established sites that don't meet the needs of tomorrow's "mega data centers" will spur the growth of a more-federated, hub-and-spoke data center topology for most large user enterprises (0.8 probability). Therefore, enterprises with multiple, global data centers will need to develop a holistic strategy for their site topologies during the coming decade. Trends Driving the Return of Large Centers Not long ago, large data centers were on the decline. In decades past, the rise of distributed computing and other trends drove the decline of the large data-processing sites that characterized the era of mainframe dominance. Now, however, data centers are rising in importance, and Gartner expects this trend to continue during the next decade (see Figure 1). Publication Date: 4 August 2006/ID Number: G00142079 Page 2 of 6
Figure 1. Center Topology Evolution Physically Integrated, Logically Centralized (Hub-and-Spoke) Center Influence Physically Centralized, Logically Centralized Physical Consolidation Physically Decentralized, Logically Decentralized (Client/Server, Distributed) Source: Gartner (August 2006) 1980 1990 2000 2010 There are a number of trends leading to the new growth of the influence of large data centers. They include: Server rationalization Users are looking to gain financial and operational benefits by rationalizing their servers and moving them into a centrally managed physical location. The net financial benefits are limited in the short term, but it makes sense to better manage hardware sprawl. Focus on cost containment The need to cut IT costs during the past few years has led many organizations to examine their data center real estate portfolios. As a result, many have reduced the number of sites by placing more equipment into the larger sites. The need to improve security and business continuity With a focus on security, business continuity and corporate governance increasing, many enterprise planners are concluding that a centrally managed data center environment with appropriate technologies and procedures provides the optimum level of protection. Growth in hardware requirements The past few years have seen a fairly constant increase in server deployments, and a more-exponential growth in storage hardware. Although the size of the hardware has grown at a lower rate than the number of units, there has been a net increase in data center space requirements. Delivering software as a service As enterprises start to deliver more packaged software over the Web in a seamless manner, there will be increasing demand for huge arrays of servers and storage boxes that will need to be hosted in mega data centers. This is especially true for companies that mainly do business over the Web, such as Amazon.com and Google. Publication Date: 4 August 2006/ID Number: G00142079 Page 3 of 6
The Emerging Center Topology The above factors would naturally lead to the development of a single, very large central site or several large sites per major geography such as North America; Europe, Middle East and Africa (EMEA); and Asia with as much technology packed into them as possible. However, due to issues of network bandwidth and the power and cooling available in any single, legacy data center a natural limitation will be reached. Moreover, in certain regions, such as Europe, cultural conflicts, language differences and legal issues surrounding data retention will continue to limit the level of site consolidation per geography. In addition, internal politics in many large companies will also create a natural limit on the number of physical data center sites that can be closed. As a result, the topology of tomorrow's data centers will be dictated not only by the operational and technological efficiencies that drive lower service delivery costs using the smallest possible number of data centers, but also by the practical realities of the enterprise's need to consider organizational, political and cultural issues, and its desire to leverage the IT investments made during recent decades. Gartner believes this will lead to the emergence of a new data center topology that comprises logically centralized and physically integrated data centers with spoke locations (see Figure 2). Figure 2. The Emerging Topology for Tomorrow's Mega Centers Center 2 Remote Location 1 Center 1 Center 3 Center 1 Center 4 Center 5 Mega Center Remote Location 2 Remote Location 3 Different Operational Processes; Similar-Size Centers Standard Operational Processes; Mega Center Hub With Remote Spoke Locations Source: Gartner (August 2006) There are four key tenets to this emerging topology: Publication Date: 4 August 2006/ID Number: G00142079 Page 4 of 6
1. Maximize the size and functionality of the large hubs, and possibly create centers of excellence. The aim is to have the least amount of redundancy of infrastructure across multiple sites by concentrating resources into the smallest number of sites and by making those sites as large as possible. Exploiting economies of scale in this way will result in the most cost-effective deployment of physical assets, such as server farms, storage devices and monitoring tools. As a secondary step, certain mega data centers may be designated as centers of excellence for certain functions. For example, one Gartner client designated a data center in the Frankfurt region of Germany as an SAP center of excellence because of its proximity to SAP's headquarters and the concentration of internal SAP skills. 2. Keep the remote/spoke locations as light as possible in terms of infrastructure and application development. With improving network bandwidth, organizations should try to limit the amount of core infrastructure at the remote locations. In many cases, an architecture that uses thin-client devices may be appropriate. management, backup and recovery may require local infrastructure, but should be managed centrally as much as possible. Application life cycle management should be carried out at the larger hubs to contain people and technology costs. 3. Use the latest virtualization and replication technologies to improve asset utilization and disaster recovery. Emerging virtualization technologies, particularly in the server and storage domains, offer the best asset utilization potential. These technologies can also improve operational processes, such as change management and capacity planning, and can enhance core disaster recovery and business continuity capabilities. replication will continue to involve a trade-off between the cost and risk benefits of synchronous and asynchronous modes. Spoke sites should be considered staging locations for synchronous data. 4. Ensure that all sites conform as closely as possible to a standard set of operational service levels and use standard service management tools. Whatever the evolving topology, users should strive to run all sites on a standardized set of operational service delivery processes. These may be modeled on an industry practice, such as the IT Infrastructure Library, or developed in-house. Also, a standard set of tools should be used, particularly for trouble ticketing, asset management and change management. By operating all sites in this manner, IT service delivery costs will be limited to an optimal level, and user expectations will be kept consistent. Moreover, such a method also makes it easier to deploy new technologies, such as virtualized servers, and to introduce changes to optimize business processes. Recommendations Enterprises with multiple data centers per major region need to develop a holistic strategy for their site topologies for the next seven to 10 years. The focus should be on developing a smaller number of large data centers, and having spoke sites with limited infrastructure and people resources. To optimize costs, all sites should conform to well-documented operational and service delivery processes. Publication Date: 4 August 2006/ID Number: G00142079 Page 5 of 6
REGIONAL HEADQUARTERS Corporate Headquarters 56 Top Gallant Road Stamford, CT 06902-7700 U.S.A. +1 203 964 0096 European Headquarters Tamesis The Glanty Egham Surrey, TW20 9AW UNITED KINGDOM +44 1784 431611 Asia/Pacific Headquarters Gartner Australasia Pty. Ltd. Level 9, 141 Walker Street North Sydney New South Wales 2060 AUSTRALIA +61 2 9459 4600 Japan Headquarters Gartner Japan Ltd. Aobadai Hills, 6F 7-7, Aobadai, 4-chome Meguro-ku, Tokyo 153-0042 JAPAN +81 3 3481 3670 Latin America Headquarters Gartner do Brazil Av. das Nações Unidas, 12551 9 andar World Trade Center 04578-903 São Paulo SP BRAZIL +55 11 3443 1509 Publication Date: 4 August 2006/ID Number: G00142079 Page 6 of 6