Sizing Your NT RAID ARRAY
- Obtaining Maximum Performance While Adding Storage
- Getting the Big Picture
- Detecting Single-Disk Bottlenecks
- Selecting Disk Counters
- Detecting RAID Bottlenecks
- Sizing Additional Disk Capacity for RAID Arrays
- Determining How Many Disks to Add
- Estimating Required Additional RAID Performance Capacity
- Disk Storage Capacity vs. Disk Performance Capacity
- Meeting Your Storage and Performance Needs
Obtaining Maximum Performance While Adding Storage
How do you know where to start when you need to add storage capacity to your existing Windows NT solution? You can't simply add more disk space and expect to improve performance. As you increase your storage capacity in an enterprise environment, you need to be able to detect bottlenecks in your RAID subsystems, know which RAID levels to consider, and know how to size your RAID arrays according to current and future performance requirements.
If you're new to RAID or just need to brush up on your technology, see Table 1 in this article for a comparison of RAID levels and definitions, or go to the RAID Advisory Board Web site for an extensive RAID review. The disk subsystem is one of the most flexible resources that you can configure in NT. How well you design your disk subsystem can drastically influence NT's overall performance.
Table 1
RAID Levels and Definitions
RAID Level |
RAID Definition |
RAID 0 Disk Striping |
RAID 0 stripes disk activity across two or more disks. This logical layout provides the advantage of better performance for read, write, random, and sequential environments. The trade-off is that RAID 0 doesn't provide any fault tolerance; if you lose one disk in your array, you lose the data for the entire array. You can increase the number of disks in a RAID 0 environment and improve the random I/O performance. |
RAID 1 Disk Mirroring |
RAID 1 mirrors disk activity across two or more disks. This logical layout provides for better read performance than one disk (especially in a multiuser environment) but lower performance in a write-intensive environment. RAID 1 provides complete data redundancy, even if you use only two disks. However, this redundancy lowers the capacity of a RAID 1 mirror by 50 percent. For example, if you have two 9GB disks in a RAID 1 mirror, you can use only 9GB of storage space. |
RAID 5 Disk Striping |
RAID 5 stripes disk data with parity information across three or more disks. This logical layout provides for better read performance than one disk (especially in a multiuser environment) but significantly lowers performance in a write-intensive environment. This RAID level provides fault tolerance through the use of parity information, allowing for the loss of one RAID 5 member disk without the loss of any data. This redundancy lowers the capacity of a RAID 5 striping with parity by a factor of 1/(number of member disks). For example, if you have three 9GB disks in a RAID 5 array, only 18GB is usable storage space. |