Requested data protection

The requested protection of data determines the amount of redundant data created on the cluster to ensure that data is protected against component failures. OneFS enables you to modify the requested protection in real time while clients are reading and writing data on the cluster.

OneFS provides several data protection settings. You can modify these protection settings at any time without rebooting or taking the cluster or file system offline. When planning your storage solution, keep in mind that increasing the requested protection reduces write performance and requires additional storage space for the increased number of nodes.

OneFS uses the Reed Solomon algorithm for N+M protection. In the N+M data protection model, N represents the number of data-stripe units, and M represents the number of simultaneous node or drive failures—or a combination of node and drive failures—that the cluster can withstand without incurring data loss. N must be larger than M.

In addition to N+M data protection, OneFS also supports data mirroring from 2x to 8x, allowing from two to eight mirrors of data. In terms of overall cluster performance and resource consumption, N+M protection is often more efficient than mirrored protection. However, because read and write performance is reduced for N+M protection, data mirroring might be faster for data that is updated often and is small in size. Data mirroring requires significant overhead and might not always be the best data-protection method. For example, if you enable 3x mirroring, the specified content is duplicated three times on the cluster; depending on the amount of content mirrored, this can consume a significant amount of storage space.