May 15 ’09
Making IT Greener: Mainframe Storage Resource Management
IT organizations and the companies they serve are increasingly concerned about energy costs and environmental impact. Electricity use by data centers worldwide doubled from 2000 to 2005, representing an aggregate annual growth rate of 16.7 percent per year, according to a report titled “Worldwide Electricity Used in Data Centers” by Jonathan G. Koomey. This means IT decision-makers must start getting serious about bringing energy consumption under reasonable control.
Meanwhile, the mainframe’s unique attributes are making it more important than ever for today’s information-driven enterprises. Approximately 70 percent of organizations and governments are running critical applications on mainframes, according to a report from Butler Group titled “The King Is Dead—Long Live the Mainframe.”
While the mainframe is inherently a highly cost-effective computing platform, expanding DASD and tape storage requirements can drive 10 percent or more of a data center’s overall increases in energy consumption. Storage managers can therefore play a significant role in making IT greener. In doing so, they will face various challenges such as deciding how much of their constrained budgets they should allocate to acquiring new, energy-efficient green storage devices—or determining whether there are less costly ways to support green IT initiatives using existing tools and resources.
New Equipment vs. Better Practices
In his 2008 report, “Storage Panorama,” Fred Moore (president of Boulder, CO-based consulting firm Horison Information Strategies) predicted that storage vendors will push energy-friendly technology solutions more heavily in the coming years. However, these solutions typically have considerable acquisition costs and don’t address the issue of energy consumption by existing equipment. It’s also unclear that acquiring new hardware is the best thing to do in a down economy, especially given that green storage technology will likely become even more energy-efficient and prices will probably fall considerably in just a few years.
A less costly approach may be to make better use of existing Storage Resource Management (SRM) capabilities. With SRM, storage managers can effectively classify storage, more intelligently migrate and archive data sets, and even move less frequently used files to tape media that doesn’t require energy on an ongoing basis. SRM supports green IT initiatives by providing storage managers with greater visibility into their current storage environments, so they can lower energy consumption by improving utilization.
SRM makes it easy for technicians to pinpoint available storage space and automate efficient restructuring of mainframe storage resources. Besides cutting energy consumption, this approach also reduces capital infrastructure costs and ongoing storage ownership burdens.
Because SRM is implemented on the host system, it doesn’t require the purchase of additional hardware. In fact, in addition to deferring the acquisition of expensive new hardware, SRM can actually allow for the removal of unneeded hardware already in place.
Driving Down Energy Costs With SRM
Here are some key points storage managers should consider when applying SRM to support green IT initiatives:
1. Use space management techniques to set goals for efficient asset utilization:
• Higher asset utilization of storage hardware is essential for reducing energy consumption. Space management is an important SRM function that dramatically improves utilization. In fact, in large z/OS installations, an improvement in utilization of just a few percentage points can help recoup several terabytes in wasted disk space.
• Over time, smart space management policies can help an IT organization avoid the acquisition of multiple, additional disk subsystems. Utilizations in the 70 to 75 percent range are fairly typical. However, SRM can yield utilizations of 85 percent or more. This differential can significantly affect the size and timetable for infrastructure purchases.
Figure 1 shows the tracking of storage capacity on the floor against usage. The goal is to raise utilization to reduce costs and environmental impact. The “capacity (red) line increases, even though the usage (green) line doesn’t converge. This indicates storage acquisition is outstripping requirements.
Figure 2 shows another example of tracking business unit or application usage. Space allocated, but not used, is shown over time. Challenge data owners to improve efficiency and only request storage capacity that’s needed.
2. Accurate forecasting based on history and trending can prevent unnecessary purchases:
• No IT organization can afford to have empty drives spinning for no actual purpose. SRM helps eliminate this type of inefficiency by providing both detailed historical measurement functions and intelligent trending tools that enable accurate predictions of future space requirements. This helps organizations purchase storage configurations of the correct size and lets them continually update their forecasts to stay on top of changing conditions that might affect future storage needs.
• Space usage data can be collected for all storage groups, disk subsystems and/or applications to provide all stakeholders with the forecasting data they need. SRM technology should be centralized and made available to all appropriate IT staff. It may also be useful to make space management reports available on intranet sites or via email. This approach also ensures the entire storage environment isn’t repeatedly scanned by multiple users and departments, which can be a drain on CPU resources.
3. Carefully coordinate the acquisition and decommissioning of DASD subsystems to minimize the amount of time they both need to be simultaneously powered and cooled:
• Hardware vendors traditionally supply disk subsystems with a free maintenance period. At the end of this period, financial incentives are usually offered to replace disk subsystems with new models that provide increased capacity and performance. Typically, this new hardware is more energy-efficient, generates less heat per terabyte, and uses less floor space per terabyte.
• Migrating data to new hardware is a common occurrence. It makes sense to use SRM and sophisticated DASD migration utilities to optimize these migrations, thus moving the data as quickly and efficiently as possible. This avoids extended periods where both the old and new disk subsystems are consuming resources.
4. Maximize the efficiency and utilization of cartridge robots:
• Tape hardware can be another significant consumer of power, cooling, and floor space. Cartridge robot devices can store thousands of cartridges for quick access. SRM technology should be used to manage these robots— tracking usage and automatically ejecting infrequently used cartridges.
• SRM also should be used to track scratch tape availability and free-cell counts to provide history and forecasting information. In this way, SRM can be used to keep to a minimum the number of cartridge robot devices installed and delay the purchase of additional robots as long as possible.
5. Use virtual tape to reduce the number of physical tape drives and cartridges:
• Virtual tape helps deliver a higher ROI by stacking many virtual volumes on a single physical volume. That means fewer physical tapes to buy, transport, and store.
• Virtual tape can be hardware-based or software-based. Software virtual tape solutions allow for the use of less tape hardware and media, thus reducing costs, improving tape performance, and improving utilization.
• SRM can be used to simplify management of virtual tape deployments by automating and centralizing administration to help maximize existing resources, achieving higher utilization and optimization of the virtual and physical media.
IT organizations can significantly reduce the energy consumption of their data centers by effectively leveraging SRM capabilities to make optimum use of their existing storage infrastructure, defer acquisition of new hardware, avoid acquiring excess capacity, and limit the amount of time spent simultaneously supporting old and new systems. To achieve these benefits, storage managers should ensure that SRM capabilities and policies are appropriately applied across the full spectrum of mainframe storage infrastructure—including DASD, tapes, robotics, virtual tape systems, and tape management systems.