News Stay informed about the latest enterprise technology news and product updates.

RAID modes on a 1U server

Here’s how RAID technology works to provide data resilency.

For decades, the use of RAID (Redundant Array of Inexpensive Disks l1) technology has been an important part of storage resilience. The goal is to combine multiple disks into a single logical entity where the combination of disks can provide a level of redundancy for stored data. Although RAID is often associated with formal storage systems such as a SAN, it can also be deployed on 1U servers with multiple local disks installed. There are numerous ways to combine disks. Each approach is called a "RAID level" and offers unique tradeoffs that system planners should consider, so let's look at a few of the most popular RAID levels available for 1U servers.

RAID 0 only provides block-level striping. This enhances the apparent performance of the RAID group because the data is interleaved across multiple disks and more spindles are spinning simultaneously. This RAID level provides no data protection or resilience. RAID 1 will directly mirror two disks. The data on one disk is simply copied to a second (and sometimes a third if additional resilience is desired). If the first disk fails, the second disk takes over. The disadvantage here is that you need twice the number of disks to store the same amount of data.

RAID 5 offers block-level striping with distributed parity. Data is interleaved among multiple disks and parity information is also calculated and distributed across the drives. If one drive fails, its contents can be reconstructed on the fly from the parity data on the other disks. The RAID group takes a performance hit until the failed drive is replaced and its contents fully reconstructed. RAID 6 offers block-level striping with double-distributed parity. It's similar to RAID 5, but the dual-parity approach protects the RAID group from two simultaneous disk failures. For example, if one drive fails, the RAID 6 group is still protected from a second disk failure until the first disk is replaced. Some vendors implement double-parity RAID differently, so pay attention to specific server requirements or limitations with RAID 6.

RAID levels can also be combined to mix tradeoffs. For example, RAID 10 combines RAID 1 (mirroring) and RAID 0 (striping) to provide performance and resilience without the use of parity. RAID 50 combines single-parity disk protection with mirroring, while RAID 60 combines double-parity disk protection with mirroring. Both effectively duplicate each RAID 5 or RAID 6 group for exceptional resilience but demand twice the disks.

One last note:  When implementing RAID on a 1U server, include a battery backup module. Disk writes are normally conveyed to cache memory on the RAID controller before being committed to physical disks. If system power fails before the writes occur, cache data can be lost -- resulting in potentially catastrophic data loss to the RAID group. A battery backup protects RAID controller cache and ensures that when power is restored, cached data will be committed to disk properly. Some organizations may opt to rely on uninterruptible power supply backup power to protect the server (and RAID controller) and orchestrate a controlled server shutdown when utility power is lost.  


Dig Deeper on Server hardware strategy

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.