It s practically a cliché, but - PDF

THE STATE OF DATA CENTER EVOLUTION By Andrew Mazer It s practically a cliché, but that s because it s true: An organization s data is one of its most valuable assets. Today, robust analytics tools enable

Please download to get full document.

View again

of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.

Real Estate

Publish on:

Views: 8 | Pages: 7

Extension: PDF | Download: 0

THE STATE OF DATA CENTER EVOLUTION By Andrew Mazer It s practically a cliché, but that s because it s true: An organization s data is one of its most valuable assets. Today, robust analytics tools enable businesses to take advantage of the full potential of customer information to grow revenue, improve services, reduce costs and efficiently target potential customers. Other data, such as financial records, repositories, transactional records, contracts and R&D have obvious strategic value and must be protected. To safeguard data, organizations must ensure that their data centers are wellpositioned to grow and evolve as business needs require. Yet data center administrators are facing an array of challenges, from the exploding volume of data that organizations generate and consume to space constraints, power and cooling concerns, and the need to protect against physical threats. Even the availability of adequate power is becoming a challenge in some locations. Executive Summary :: Most IT executives say that some of their organizations data centers will max out of power, cooling or space by the end of 2014, according to new research from UBM Tech and InformationWeek Marketing Services. The study, conducted in April 2013, takes an in-depth look at how IT decision makers rate the challenges they face as they evolve data centers to handle fastgrowing volumes of data. Among the key findings: Only 17 percent are actively planning or undertaking to build a new data center. However, over the next two years, 29 percent of respondents say they plan to perform a major upgrade on one or more data centers and 20 percent would consolidate data centers into fewer facilities. As few as 18 percent plan only to maintain their current data centers without significantly upgrading hardware. Rather than investing capital in new buildings, respondents are upgrading equipment, consolidating and refreshing hardware. More than half of the IT execs are actively considering moving forward with server consolidation, blade deployment, virtualization and outsourcing servers/storage, and installing more energy-efficient equipment. Sponsored by UBM Tech and InformationWeek Marketing Services surveyed 100 business technology decision makers at North American companies with more than 500 employees to gain insight into the state of their data centers, the challenges they face and the strategies they re using or planning to use to monitor and manage their IT infrastructure. Data Centers Are Aging Faster than Anticipated: Colo Is Gaining Momentum More than half the IT decision makers surveyed (57%) say that the state of their data centers power, cooling and space is fair, serious or urgent meaning that they anticipate running out of capacity before the end of As they approach this brick wall at light speed, these organizations will need to quickly identify and deploy solutions that enable them to meet service-level agreements and other commitments. Thirty-one percent of IT pros flagged aging facilities as their chief problem today, followed by concerns about running out of processing capacity () and fear of losing data (15%). This is a notable change from priorities just five years ago: In a 2008 Siemens-sponsored study of data center energy efficiency, aging facilities were much less of a concern only 15 percent cited that as their biggest problem. Back then, the big worries were lack of processing capacity () and the risk of losing data (25%). To respond to the threat of running out of capacity, nearly 30 percent of IT executives said they have moved or are currently moving to a colocated facility (see Figure 1), and approximately another 20 percent said they are considering doing so. A lot of businesses are outsourcing to colocation facilities, especially after Hurricane Sandy, when many lost power and connectivity for days, says Kelly Quinn, research manager in the Data Center Trends and Strategies group at IDC. They install their own equipment and essentially create their own data center in another facility, which provides power, connectivity, security and support. One survey participant, data center architect Johnny Martin of Independence Blue Cross in Philadelphia, said that his company had placed both its production and disaster recovery data center in colocated facilities: Colo gives us cost savings, especially when we compared the depreciation and CapEx to the cost of new construction on stand-alone property. It also was a great vehicle for improving efficiency, because we were able to accelerate turnaround times by 80 percent. With colo, you re not beholden to time-consuming internal reviews by other groups, and you have much tighter control over the entire environment. Another approach to extending capacity is to go modular. More than a quarter of IT executives said they are planning Figure 1: What are your organization's plans when it comes to the following initiatives as it strives to meet its growing compute demands? Completed within last two years Actively planning/ Currently undertaking Use blade technology 33% 32% Consolidate servers in existing data center(s) 38% Build a new data center 23% Add a data center infrastructure management software solution 18% 20% Implement virtualization/outsource server and/or storage to third-party cloud providers 30% Put more energy-efficient equipment in place 32% Upgrade with electrical improvements 12% 21% Upgrade with mechanical improvements 11% Expand size of existing data center 10% Pursue green certifications for current or future data centers (e.g., LEED, Energy Star) 9% Lease colocation space to host new equipment 7% 22% Adopt modular technologies 7% Enlist third party to develop and execute efficiency and/or sustainability improvements 3% to install, or had already installed, a modular data center. We ve seen a number of companies set up modular additions to their data centers, which are not only relatively inexpensive, but also can be deployed quickly, says Quinn, The only catch is that you need a place to put it either on the roof or somewhere near the data center. Also, you can t expand the capacity of a container; they are scalable only by adding new ones. 2 Contending with Escalating Costs Many companies have had to postpone large capital investments in IT infrastructure due to economic constraints over the past several years. The good news is that 58 percent of companies will have an increased IT budget this year, and 25 percent of those will see their budget rise by more than 10 percent. Still, the expenses of running a data center are also increasing. Nearly two-thirds of respondents said their data center costs have risen over the past few years. And half of those surveyed said they expected that energy consumption in their data centers would rise over the next three years. Power and cooling costs are driven not only by utility costs, but also by the deployment of high-density blade servers, which pack impressive computing power in a small footprint. Blade deployment is the most common initiative that companies have taken to meet increased computing demand, according to respondents, followed by consolidating servers in existing data centers and building a new data center. Although less than one-fifth of respondents said their data centers had serious or urgent issues with power, cooling and space, 40 percent said that at least some facilities would reach capacity by Today s high-density blades are at least partially to blame; despite their small footprint, current-generation blade servers can be power hungry, often consuming five times or more per rack than previousgeneration hardware. 1 When asked what the biggest technical challenges in the data center were, most IT executives said future-proofing Methodology In April 2013, UBM Tech and InformationWeek Marketing Services conducted an online survey on behalf of Siemens on The State of Data Center Evolution. UBM Tech collected data through an online survey from 100 IT business technology professionals involved with data center management or decision making at companies with 500 or. About half of respondents were IT executives (CIO/CTO or VP of IT) or directors/managers of IT. The greatest possible margin of error for the total respondent base (N=100) is +/- 9 percentage points. UBM Tech was responsible for all programming and data analysis. These procedures were carried out in strict accordance with standard market research practices. their data centers to meet growing and changing demands (57%), followed by reducing complexity and improving operational efficiency (50%), and avoiding expensive power interruptions (45%). Other frequently cited challenges were managing assets, tracking changes to the data center environment, and attaining energy efficiency goals (see Figure 2, p. 4). Tracking Energy Consumption Growing in Importance IT execs are clearly interested in reducing energy costs. More than three-quarters (78%) of respondents consider energy consumption somewhat, very or extremely important in planning for expanded computing capacity. Most companies are using a variety of tools to measure the power and cooling impact of adding servers. The most common approach is employing dedicated metering to track energy used by cooling systems and IT systems (23%), closely followed by capacity management tools to avoid excessive power and cooling demands (22%). Slightly fewer respondents said they used multiple power meters to measure power usage over time and carefully considered where to locate new servers to minimize cooling demand (see Figure 3, p. 4). Although many IT pros are closely tracking energy usage in their data centers, 42 percent of respondents said they seldom or never monitor key performance indicators (KPIs) of energy efficiency in the data center. Twenty-two percent said they track KPIs continuously, and 16 percent said they monitor hourly or daily. About one-third of respondents said that energy costs comprise between 20 and 30 percent of their data center s total cost of ownership (TCO), but almost half said they had no idea of the relationship. The rest were evenly divided above and below that range. Similarly, almost two-thirds of the IT pros surveyed do not know the power usage effectiveness (PUE) rating of their largest data center, and more than half are not sure of their organization s PUE goals. Only 9 percent of respondents stated that their largest data center had a PUE rating of less than 1.5, indicating that there is plenty of room for improvement. More companies measure PUE at a simple Category 1 level (18%) than at the more detailed Category 2 or 3 levels (12%). About one-quarter have set a goal to either maintain or reduce their PUE rating. Despite the unfamiliarity that many IT decision makers have regarding actual power-usage numbers in their data centers, there is a widespread awareness of the importance of managing energy usage and trying to reduce it wherever possible. For example, as they make their data centers more environmentally friendly, IT execs anticipate the largest 3 The State of Data Center Evolution Figure 2: Thinking about your organization's data centers, rate how challenging the following are to your organization. Future-proofing to meet growing and changing demands 57% Reducing complexity and improving operational efficiency 50% Avoiding expensive power interruptions 45% Managing our assets knowing what we have, where they are located, and their maintenance state 44% Keeping track of changes to the data center and its assets 43% Attaining our energy-efficiency goals 41% Delivering adequate cooling, economically and efficiently 38% Efficiently defining and scheduling service orders and tasks to reduce downtime and equipment failure Ensuring uptime with real-time monitoring, alarms and event notification Predicting the outcome of power and cooling decisions before IT hardware is installed Note: Percentages reflect combined scores of 5, 6 or 7 on a 7-point 20scale where 30 7 is very 40challenging. benefits will be lower energy costs and better compliance with regulations. More than 40 percent say their companies will take further steps toward making their data centers greener, and an equal number say they would like to go in this direction. Going Green in the Data Center Is Seen as Business Imperative Going green not only reduces our utility bill, it also helps us become a leaner, more efficient organization, says Martin of Independence Blue Cross. We re simplifying as much as possible, consolidating our distributed servers and storage platforms, eliminating tape silos and so on. We re trying to tune the physical environment as finely as we can. Still, employing energy-saving technologies can require an initial investment that many IT budgets cannot 4 Figure 3: What processes do you use to evaluate the power and cooling impact of adding servers? We track energy used by our cooling infrastructure and IT systems on separate meters 23% We use capacity management tools to avoid excess power consumption and cooling demand 22% We use multiple online power meters to measure power usage over time We carefully consider the best location to install new servers to reduce cooling demand 44% We use a DCIM system to collect and analyze information about the data center's resource use and its impact on power consumption We measure and budget power and cooling capacity at the rack level We simply anticipate spending about 66 cents (or another amount) for power/cooling for every dollar we spend on new hardware 14% We use computer modeling to visualize the impact of new equipment on power and cooling 4% Note: Multiple responses allowed accommodate. So, rather than invest in costly projects such as installing solar panels, companies are going after low-hanging fruit. Most companies today that are going green are primarily interested in cutting their power and cooling expenses. They are looking for short-term benefits and use approaches like hot aisle-cold aisle assignments and perhaps ultrasonic cooling, says Kelly Quinn of IDC. However, some data center providers, in an effort to keep costs low for their customers, are embracing more sophisticated technologies such as ambient cooling using outside air and fuel cells to create electricity. The vast majority of IT decision makers have taken proactive steps to reduce power usage in their data centers (see Figure 4, p. 5). The most common techniques cited are optimizing air handling to be as efficient as possible (51%), deploying power management tools (41%), cold aisle-hot aisle containment () and detailed power monitoring (35%). Companies planning to further reduce their electric bill this year are planning to limit consumption by idle resources The State of Data Center Evolution Figure 4: Which of the following practices have you employed in your data center to reduce power usage? Efficient air handling Using power management tools 41% Cold aisle/hot aisle containment Detailed power monitoring 35% Modular data center design Raising inlet air temperatures 21% Liquid cooling 18% Using variable frequency drives on chillers, air handlers, pumps DC power Using airside/waterside economizers 10% 18% Note: Multiple responses allowed 51% Figure 5: What are your goals for managing data center power consumption this year? Reduce idle resource energy usage 40% Improve power usage monitoring 34% Reduce energy consumption by the support infrastructure 30% Improve the efficiency of power delivery to servers 29% Buy and manage hardware that is able to consume power proportional to its required level of performance Improve our ability to anticipate server workloads Perform reviews to identify opportunities to improve power utilization Add additional incoming power capacity 14% Analyze, perform and execute power studies 14% Note: Multiple responses allowed (40%) and the support infrastructure (30%), improve power The UBM Tech survey asked IT experts about their use of usage monitoring (34%) and make power delivery to servers Data Center Infrastructure Management (DCIM) solutions, an more efficient (29%). About a quarter of respondents said integrated set of tools that enable IT administrators to remotely they planned to buy hardware that dynamically adjusts monitor and manage hardware assets, electrical 51% power delivery, 20 power usage 30 to the level 40 of performance 50 required 60 by a given equipment and facility cooling, fire and life safety systems, and workload (see Figure 5). building security. In the 2008 Siemens study, organizations were using While only 15 percent of IT experts said they were using different strategies to reduce energy usage. Back then, a DCIM solution, almost one-third said they were using virtualization was the most common project, cited by 45 some DCIM tools, and 22 percent indicated they would be percent. Installing more energy-efficient hardware and interested in deploying DCIM capabilities. Figure 6 (p.6) upgrading air conditioning and cooling systems were shows which DCIM functions respondents are using and distant followers, each flagged by 19 percent. how they rate the importance of each. An effective DCIM system can intelligently integrate Managing Data Center Infrastructure with information from IT and facility assets, workflows and DCIM Becoming a Necessity work orders, and execute what-if analyses to enable IT Any plan to optimize a data center to handle growth would administrators to optimize capacity planning and energy use. be incomplete without addressing management tools. IT For example, DCIM solutions with computational fluid administrators need detailed visibility and control over all aspects dynamic (CFD) modeling enable IT to create a virtual of data center operations so they can monitor the infrastructure data center that predicts and visualizes the outcome of and identify potential problems before they escalate. power and cooling decisions before hardware is installed. 5 DCIM solutions also address asset management through all life-cycle stages from design, planning and installation to operation and migration. This capability enables IT to conduct a what-if analysis before installing, moving or changing any equipment, enabling better analysis of how a different layout or scenario will affect operations. Real-time monitoring and notification capabilities enable IT administrators to take proactive steps to minimize downtime and react effectively when necessary. DCIM helps create a holistic view of all your resources so you can plan better, says IDC data center expert Jennifer Koppy. It helps companies gain clarity into all the resources in their IT environments and identify the idle or underutilized assets that aren t doing productive work. Then you can ferret out these zombie servers and clean up the environment. With workflow, change and document management capabilities, a DCIM solution allows stakeholders in data centers to work in closed-loop processes that facilitate the accurate and rapid performance of service tasks. The most common DCIM functions being used are power monitoring, fire and safety systems, asset tracking, real-time monitoring, device alarming and UPS/emergency generator controls, and physical/perimeter security systems. These same functions were considered important by more than two-thirds of all respondents. DCIM makes managing a data center much easier. I can splice any metric I want to track into it, so it s easy to monitor hardware and set up alarms. We can get cabinet reports, branch circuit monitoring, and real-time views of what cabinets and free-standing equipment are drawing, says Martin. We use power tracking to verify our monthly invoices from the colocation providers and shrink our energy use. Having the ability to home in on power usage by rack, cabinet or freestanding mainframe really gives us an edge. Just as importantly, DCIM enables data centers to perform more efficiently
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks