|
|
|
NETWORK DEMANDS NEW PARADIGM FOR DISK STORAGE SUBSYSTEMS
Efforts among major systems, storage subsystem, and disk
drive vendors to invent a Network Attached Secure Disks (NASD) architecture
will result in a new storage subsystem paradigm by the late 1990s, according
to recently published research from International Data Corporation (IDC)
of Framingham, Mass.
Serial Storage Architectures (SSA) and Fibre-Channel (FC) interconnects
with high bandwidth and network packet protocols will be enabling technologies
for directly attached network disks. This new technology will allow
disk and tape subsystems and hosts distributed throughout a building
or over a campus to interact as if they were a tightly coupled system.
The NASD concept will displace traditional store-and-forward data transfers
through a server, while increasing client/server scalability by an order
of magnitude. Data will move between client and disk at a fast disk
subsystem rate. “Although the NASD model looks appealing, there
are some very important questions that still must be answered,”
said Robert Gray, IDC Research Manager, Storage Subsystems. These include
the following:
• What functions should be designed into the disk and storage
subsystem?
• How does the NASD paradigm fit into other styles of computing?
• What are the expected network access times (i.e., latency) ?
• How much network and file system knowledge should be in storage
components?
Powerful, flexible interconnects such as SSA and FC, which can handle
the strain of a more robust network connection will enable the shift
to new storage subsystem and networking products. Servers, however,
will remain limited in scalability. This results from a storage-to-server-to-client
data/program copy problem.
<BACK
|
|
|
