The process of raw device mapping (RDM) is used in the context of storage area networks (SANs) to help provide segmentation and order for these large networks. In storage area networks, virtual machines and other components are clustered together to provide robust storage capability.
Companies use raw device mapping to aggregate virtual machines under a logical unit number or LUN. The LUN will “own” a range of virtual hard drives that are part of the storage build. By assigning a virtual machine to a LUN, engineers are allowing mapping files to act as a proxy for the device.
In some ways, the use of raw device mapping and virtual machine file systems (VMFS) are a bit of an advancement from traditional redundant array of independent disks or RAID setups. Storage is getting a makeover in an age where storage is increasingly linked to converged and hyper-converged systems, and where it plays an important role in virtualized networks and other high-tech, complex architectures.
Raw device mapping is an alternative to a virtual machine file system. Both of these technologies can allow engineers to cluster virtual machines in specific ways. One specific use of raw device mapping is to accomplish guest operating system clustering. In general, raw device mapping helps to assign a role in a place for your virtual machine inside the system.
In talking about why a company would choose raw device mapping over a VMFS or datastore, experts sometimes cite the use of user-friendly persistent names – in other words, raw device mapping helps to give clearer and more transparent names to components. Other benefits include the use of snapshots, where engineers can pull up virtual machine snapshots on mapped volumes. Raw device mapping can also be used to migrate a virtual machine with vMotion. Then there’s dynamic name resolution, where raw device mapping stores unique information for each virtual machine. All of these features add to the utility of RDM to achieve storage results in a complex IT architecture.