Jul 7 ’09

Storage & Data Management: IT Resource Optimization and Efficiency—It’s About Time and Space

by Editor in z/Journal

What’s your definition of IT resource (e.g., server, storage, network, facilities, and software) efficiency and optimization? If it’s consolidation or boosting utilization to do more with what you have, or to do more with less, including server or storage virtualization, you would be correct. However, if you said it meant boosting performance, doing more work, storing more data, and moving more information in a smaller amount of time while using less energy and floor space to meet a given service level, you also would be correct.

For legacy enterprise environments, this goes without saying; however, sometimes what’s assumed and understood needs to be discussed. Thus, in the quest to reduce, defer, or otherwise shift or mask costs by striving toward higher levels of IT resource utilization, it’s important to keep performance, availability, and overall Quality of Service (QoS) in perspective. For example, a server or storage system may appear to be underutilized, but, in fact, be delivering a specific performance or QoS level, thus meeting its service objective; however, are the IT resources or applicable technologies being optimally used? The answer to that is it depends!

While there’s a tendency to focus on capacity (space) footprint utilization improvements, including data reduction ratios, there also needs to be a balance between and focus on performance and response time requirements. Thus, IT resource efficiency and optimization are balanced between time (performance) and space (capacity) while maintaining availability to a given service level requirement. Techniques and technologies to boost IT resource performance (time) efficiency include:

• Caching and performance acceleration techniques on a local- and wide-area basis

• Balancing data transfer rates with data reduction ratios

• Leveraging fast storage, servers and networks for efficiency

• Consolidating to faster technologies that do more work in less time using less energy

• Measuring efficiency in terms of work or activity per watt of energy or data transfer rates.

Techniques and technologies to boost IT resource capacity (space) efficiency include:

• Reducing data footprint impact across all applications and categories of data or storage

• Archiving and data management techniques for compliance and non-compliance data

• Policy-based de-duplication that balances time (transfer rates) and space (reduction ratios)

• Tiered storage, including high-capacity disks for near-line and tape for off-line and archive

• Intelligent power management for servers and storage such as second-generation MAID 2.0

• Various RAID levels to balancing performance (time), availability, capacity (space)

• Measuring efficiency in terms of idle capacity per watt and data reduction ratios.

Additional general tips to boost efficiency include:

• Examine how existing software licenses can be optimized and used more effectively

• Balance time and space to meet different application needs.

Find a balance of Performance, Availability, Capacity, and Energy (PACE) to a given application service requirement, thus aligning the applicable tier of resources to the task at hand. That’s balancing the need for time vs. the need for space for different application or data tiers, which enables you to address and fix problems instead of just moving problems.

By fixing bottlenecks, the result is improved efficiency and optimization. For example, if a storage system has a low utilization to support a performance-sensitive application, then put that application and data on fast storage. Use fast Fibre Channel or SAS disks with cache. An alternative would be to use RAM-based Solid State Disk (SSD) for read- and write-intensive data, or FLASH-based SSD for read-intensive data. Another example is to leverage fast servers to get time-sensitive work done during business hours, and then use available server performance for off-hour IT Infrastructure Resource Management (IRM) tasks, including database or file system maintenance, backups, archiving, or other compute- and resource-intensive tasks. The result is improved efficiency and optimization by more fully using resources.

The bottom line: Common sense combined with best practices can be used to balance PACE to a given service level and cost requirement. After all, it’s about time and space!