At this point in time, most customers expect to have these features in both block- and file-based storage products. In block-based products, these features have historically been between "like" systems only (systems from the same vendor), but that's been changing over the last few years to become more heterogeneous, though heterogeneity is still the exception. The features are available, but there may be operational issues that add to the complexity of a unified [or multiprotocol] storage platform.
Features, like snapshots and replication, are all based on the granularity supported by the block storage system. This is typically at a LUN level, so I can take a snapshot, such as a Copy On Write (COW) or a clone, on a specific LUN. If I'm doing a remote replication, it's typically on a LUN or a set of LUNs. This is more sophisticated with write ordering requirements between LUNs to maintain database integrity. When you move to file-based systems like NAS, snapshots and replication are handled at the file system level, which works very differently. When I combine these functions into unified storage
, the way these functions work and the way they're administered will be slightly different. There may also be operational differences depending on the storage target.
Listen to the unified storage FAQ audiocast.
Go to the beginning of the Unified Storage FAQ Guide.
06 Feb 2007