Information Technology Reference
In-Depth Information
port of very large data sets. Sufficient redundancy across multiple devices
ensures that storage remains viable even in the event of individual worksta-
tion reboot or shutdown in these distributed storage systems, although care
must be taken to avoid overwhelming network capacity.
Direct Attached Storage (DAS)
The most rudimentary storage configuration involves storage devices and
arrays connected directly to a client computing system. This includes
both internal and external devices, attached via a wide range of interface
technologies. Thin client systems lack extensive direct attached storage,
consuming remote resources for storage and operational capacity. Direct
attached storage is generally the most difficult to expand, due to hardware
constraints on the maximum number of drives that may be connected.
Network Attached Storage (NAS)
Backup and access audit control can be greatly improved by relocating
data to centralized networked storage locations and dedicated network
attached storage devices. This is the first step towards a coordinated
enterprise storage solution, and requires careful planning to ensure that
adequate resources and connectivity are present for all consuming client
systems. NAS solutions should be considered when planning migration
from client to centralized server storage if an intermediary storage solu-
tion is required before a central data entry storage system can be fully
implemented. Placing network attached storage arrays at departmental-
level data closets can begin the process of data abstraction from DAS
standalone configurations.
Storage Area Network (SAN)
Unlike network attached storage solutions, which use software and
hardware to provide networked resource access across standard network
interconnectivity, a storage area network employs specialized high-perfor-
mance network transport protocols and dedicated network connectivity to
aggregate storage and consuming systems into a more cohesive high-speed
mesh. SAN solutions allow greater control over resource allocation and
