NETWORK STORAGE SOLUTIONS
Overview
Data storage is an essential aspect of any business. Whether it's project files or expenditure reports, important data needs to be stored in a central location that is not only secure but also easily accessible to the people that need it. Network storage provides a reliable data storage system that all computers on the network can connect to in order to access and share files and folders securely and efficiently, while providing a level of protection against data loss.
RAID (Redundant Array of Independent Disks)
RAID is a method of configuring multiple hard drives into one logical unit to store data redundantly and improve drive performance. Data is distributed across the drives in one of several ways known as RAID levels, each with varying degrees of performance and fault tolerance. Fault tolerance means that your data is secure when one or more of the drives fail. Each RAID level is based on one or more of three methods: Striping, Mirroring and Striping with parity.
- Striping: Data is spread across multiple drives improving drive performance but it offers no fault tolerance.
- Mirroring: Data stored on one drive is mirrored onto another drive to provide a high fault tolerance and drive read performance.
- Stripping with Parity: Data is spread across multiple drives and a portion of each drive, or sometimes an entire drive, is reserved for generating parity (which is additional data used for recovery). This provides good fault tolerance and drive read performance.
The four most common RAID levels are RAID 0, RAID 1, RAID 5 and RAID 10 (1+0).
RAID 0 - This is the cluster-level implementation of data striping where data is evenly written across multiple drives. This increases performance because multiple drives are reading and writing data, improving drive input/output. However, the downside is that there is no fault tolerance. If one drive fails it affects the entire array and the chance of data loss is high. A minimum of two drives are required for RAID 0.
RAID 1 - This is the pure implementation of data mirroring where data is copied seamlessly and simultaneously, from one drive to another creating a replica. The downside to this is that there is a slight drag on performance because data has to be written twice and it cuts your total drive capacity in half. However, it provides a high fault tolerance because if one fails the other drive can keep working until the failed drive can be replaced. A minimum of one pair of drives are required for RAID 1
RAID 5 - This is a cluster-level implementation of data striping with distributed parity where data and parity are evenly distributed across multiple drives for enhanced performance. If one of the drives fail, the system is still fully operational and the data on that drive can be reconstructed using the distributed data and the parity data from the other drives. The downside to this is that drive performance takes a hit on systems that perform a lot of write operations. However, it still provides better overall drive performance than mirroring, but less fault tolerance.
Another benefit of RAID 5 is that it allows drives to be "hot-swappable", meaning that a failed drive can be swapped with a new drive, and the data reconstructed, without shutting down or interrupting users that are accessing the system. A minimum of three drives are required for RAID 5.
RAID 10 or RAID 1+0 - This is the most common and recommended implementation of mirroring and striping. It combines the mirroring of RAID 1with the striping of RAID 0 (the mirroring is done before striping). This provides the best drive performance and fault tolerance and allows the system to remain operational even if multiple disks fail. It is most beneficial for systems that perform many write operations. A minimum of four drives are required for RAID 10.