What is RAID 0?

RAID 0 is a permanent feature among the list of common RAID levels, even though, strictly speaking, the standard isn’t even a Redundant Array of In­de­pend­ent Disks. The core principle, i.e., redundant data storage, isn’t given with this level of hard drive coupling. In a RAID 0 network, which is always composed of at least two near or fully identical storage media, a single logical drive is created to optimise read and write access. Compared to other RAID levels, data security doesn’t improve with RAID 0.

RAID 0 is a stand­ard­ised RAID level that describes the com­bin­a­tion of two or more hard drives for the benefit of per­form­ance op­tim­isa­tion. For this purpose, all data is evenly divided into ‘stripes’ or blocks on a par­ti­cip­at­ing storage media. This is also known as ‘striping’. Let’s look at an example. Say hard disk 1 receives data block ‘A1’ during a write process, while hard disk 2 stores data block ‘A2’ at the same time. Combined, the two blocks form data record ‘A’ or a part of this data record (depending on its size). If the data is requested later, it can be read in parallel.

RAID 0 uses striping to maximise both write and read speeds. However, if a hard drive fails, the data is lost because intact hard drives only have their re­spect­ive stripes stored on them.

Par­al­lel­ised data access is one of the great ad­vant­ages of a RAID 0 array over a single hard drive. The network provides more bandwidth, and auto­mat­ic­ally increases the number of possible input and output op­er­a­tions per second (IOPS for short). However, because the use of SSDs in a RAID array comes at the expense of per­form­ance, this advantage doesn’t apply to newer storage devices. RAID 0 is geared toward the use of HDD hard disks more so than other RAID levels.

A decisive dis­ad­vant­age compared to a single storage medium is the higher risk of failure. Each hard disk in the network can fail on account of hardware or software problems, thereby causing the entire system to fail. Im­port­antly, as the number of linked data carriers grows, so does the risk of total failure.

This is com­poun­ded by the fact that such a scenario is almost in­vari­ably as­so­ci­ated with a loss of the majority of the stored data if a separate backup strategy isn’t followed. As mentioned, RAID 0 doesn’t offer any re­dund­ancy compared to other RAIDs. Following failure, striped volumes may be missing data. In­di­vidu­al files may be re­triev­able from intact storage memory in a RAID 0 system.

The strengths and weak­nesses of a RAID 0 system clearly show the type of projects for which this storage approach is suitable. HDDs linked to one another according to RAID 0 provide excellent per­form­ance sur­pass­ing that of in­di­vidu­al hard disks. Non-critical ap­plic­a­tions such as software for audio or video pro­cessing, where data needs to be read and written quickly, benefit from this approach. In this case, a RAID 0 system makes for an in­ex­pens­ive al­tern­at­ive to an SSD. As a storage setup for sensitive in­form­a­tion such as customer data or private files, the concept is not suitable due to lack of re­dund­ancy.

The lack of a backup mechanism is unique to RAID 0. All other stand­ard­ized RAID levels use methods to prevent data loss. Some ap­proaches such as RAID 1 or the com­bin­a­tion RAID 10 rely on mirroring of data: files are always stored on at least two different hard drives. Other modes such as RAID 5 and RAID 6 use parity in­form­a­tion generated during the write process in order to re­con­struct data in the event of a hardware defect.