I have NO knowledge specifically of this particular one so will ONLY reference generalities towards this concept.
Creating an external intelligent storage subsystem that attaches to a main system VIA a disk, network, or some channel interface has been used for decades.
Conceptually this does what subsystems in computers have always done. It segregates out a group of operations and provides access to these operations through an interface so that much of it's intracacies are handled by local intelligence. (Initially, intelligent serial controllers through ISA/PCI, many Ethernet controllers, Graphic cards through AGP etc.)
After all, a file server doesn't truly need to care about how data is stored, or where it resides, it really only needs to have what appears like a physical drive, and how to interface to it.
Once only available in the realm of large systems, and only using proprietary interfaces between the storage subsystem and the server, this concept/technology eventually migrated into SCSI based systems, then into IDE systems, and more recently into SATA systems.
The two most common terms that describe storage subsystems are "SAN", and "NAS". ("SAN" is generally connected through a Fiber channel to a server, and "NAS" is connected through a standard networking hardware/software transport.)
As is the case of technology in general, storage subsystems are become less expensive and more available to "average people" rather than being ONLY feasible for "large corporations".
In a good general purpose configuration I would consider the following to be true:
1 - The subsystem attaches to the computer through
a standard interface, and the volumes created
appear as standard drives to the server.
(SATA in this case.)
2 - The subsystem uses standard interfaces for the
individual drives. (SCSI in this case.)
3 - The subsystem provides for online hot swapping of
drives.
4 - The subsystem provides for online hot swapping of
drives for the purpose of RAID enlargement.
5 - The subsystem provides for migration of volumes
from different RAID levels, or from different
storage configurations transparently to
the server. (This is not intended to mean that
this will occur while the server is running, only
that the subsystem can handle this without having
to reinstall the server.)
6 - If I were to use this for a server on a network,
then I would also look for redundant, hot
swappable power supplies. (Obtain, an extra one
in case they are discontinued as they are often
proprietary.
7 - Does the storage subsystem have management
software for the operating system intended to be
used on the sever.
These are the things that I would look for in a storage subsystem.
NOTE: I have seen cheap storage subsystems that say, "Yes, drives can be hotswapped" and have had the caveat that this CANNOT be done while the RAID is "online" to the server.
A storage subsystem enables the Operating System problems to be more isolated from the storage system. The end result should be less driver conflicts, and it should be more efficient.
The intelligence providing in the storage subsytem should be able to handle the RAID functions and won't have to deal with any other activities aside from performing RAID and communicating to the server.
Data from the RAID array should "ALWAYS" get to the server at the optimum speed because there are no drivers that could be chewing up unexpected CPU cycles to slow it down.
I have found that there is a lot a variability in storage subsystem drives, optimizations, and abilities. Often it is not easy to determine completely if the one chosen is suitable for a specific task or not, until it is in use in that task.
The BEST option is to find someone that is using a particular storage subsystem in an environment that is close to what your intended use would be, and get their experiences.
Some other nice things to look for:
1 - How much RAM can be added to the storage subsystem.
2 - How many RAM slots are there and what caveats exist in using them.
3 - Can volumes be assigned caching priority and/or can volumes be allocated RAM for caching.
4 - Does the storage subsystem keep logs, is the amount of logs available programmable, how are the logs accessed.
5 - Can array sizes be changed on the fly, while the storage subsystem is online.
While a good external storage subsystem will outperform most internal controllers it seams to me that this is overkill for most "individuals". (Outperform is used in the MOST general sense, and to include all attributes including perormance, manageability, reliability etc.)
Then again it's all a matter of needs and perceived needs and ability and desire.
Edit - Changed one line.
Edit - clarified subsystem as "external". After all any RAID controller creats a subsystem within the server to a degree.
<P ID="edit"><FONT SIZE=-1><EM>Edited by jim552 on 07/29/04 03:02 PM.</EM></FONT></P>
Facebook
Twitter
Instagram