Is IBM's File Server I/O Processor (FSIOP) really the file server entree we've
been waiting for? IBM's been getting a lot of flack about FSIOP. Here are some
of the complaints:
o It's just a 486 "bag" on the side of the AS/400.
o It runs only on OS/2 LAN Server 2 and has no management graphical user
interface (GUI).
o It uses expensive AS/400 DASD for warehousing client data and programs.
o It has no hardware upgrade path. (No socket for faster future
microprocessors.)
o Etc.
Gosh! Who would buy one of these things? Well, as an IS director of a small to
medium-sized AS/400 shop, I might! Here's why:
The real cost of implementing a LAN is not only the cost of the hardware. It's
the cost of maintaining the LAN resource itself. Configuring new profiles,
making connections, integrating services, backing up files, and plain old
problem solving are all tasks that consume tremendous personnel resources. If I
bring in a new file server, suddenly I've got two sets of problems: my current
AS/400 SYSOP tasks and my new server network administrator tasks. What am I to
do? Go out and hire another person?
And what about my users? When they want access to a new set of servicesùor if
they're having problems with a current service--who will they talk to? The
AS/400 SYSOP or the network administrator? Or both?
That's the real beauty of IBM's FSIOP solution. IBM's FSIOP integrates the
functions of a SYSOP with the functions of a server network administrator. With
the FSIOP, my present SYSOP can manage these duties right now, from the same
seats where she's been sitting all along. Changes in configuration, user
profiles, and service access are managed as an integrated part of the OS/400
administration. OS/2 LAN Server maintenance is available from any green screen,
so my SYSOP doesn't have to handle two screens on her crowded little desk.
Backup and recovery flows quickly and easily into her preestablished, time-
proven routines. And when users have problems, they know immediately where to
go for help. There's no finger-pointing, no long conferences, and no crossed
communication.
Don't get me wrong: hardware costs and performance are important
considerations. But in my 25 years of experience with hardware, I know there's
always somebody with a faster, cheaper hunk of iron who's ready to sell me a
piece of the action. I don't need more hardware right now, thank you. I need
better network management tools. And so far, there's only one vendor out there
who's selling me a piece of that solution. Thanks IBM! Thanks for listening!
This is a good start!
--Thomas M. Stockwell
Tom...I'm glad to see they're letting you out of the Home to write, but get a
grip. In its current configuration, the FSIOP is, well, let me put this in
technical terminology: a boondoggle. Few "medium-sized" shops would choose to
run a 486/66-based version of LAN Server. The 486/66 is not a worthy chip; its
performance under network loads is underwhelming at best.
Also, since most folks now use PCs rather than dumb terminals (sorry, mentally
challenged workstations), there's no need for two screens on your SYSOP's
"crowded little desk." (Hint...get a bigger desk!)
The key issues on the FSIOP are that it is a bag on the side of the AS/400,
that it does use high-priced IBM memory, and that it doesn't use a ZIF socket
for the CPU. Why should people limit themselves (and their users) to a closed-
end solution like LAN Server? The world is full of bright, young faces who are
fully versed in NetWare and NT.
There's much more support for those products (due to their large constituency)
than for LAN Server. This support makes getting help, fixes, and tips on
network management much easier. Since you're going to have to train your SYSOP
on some network operating system, why not go with one that has a broad support
and user base?
You say you need better network management tools? Sheesh, with LAN Server? Now
that's a laugh! Habla NetWare? Parlez-vous NT? If you're determined, however,
to remain closed-minded about the future of LANs in an AS/400 environment, go
with the LAN Server version of the FSIOP.
It seems that workstations aren't the only things being mentally challenged
these days.
--Kris Neely
I applaud IBM for creating a product that allows the AS/400 to act as a PC file
server. The FSIOP is long overdue and an important step in the future success
of the AS/400. When you combine one of the AS/400 server models with an FSIOP
and V3R1's Integrated File System, you've got a box that can seriously compete
with other PC file servers on the market.
I had the opportunity to try out this configuration at the recent COMMON Expo,
and I was impressed. Anyone who's suffered with the slow speed of shared
folders would have to agree it's a big improvement. With an FSIOP, you can
store PC files on your AS/400 with access speeds similar to the speed of
storing them on your local hard drive. It also has some other nice features,
like the way the system automatically keeps AS/400 profiles in sync with your
LAN profiles, and the fact that you can keep your AS/400 and LAN data backed up
with a single save/restore strategy.
Sure the FSIOP has some faults. It could stand to have a faster processor,
support for user upgradable memory, and more choices of network operating
systems. But let's face it: this product is only in its infancy. I believe IBM
will address these and other concerns in the future. IBM has already made a
statement of direction to support Novell NetWare. I would expect to see other
improvements like this over the next few years.
Obviously, the FSIOP is not for everyone, but for those who want a turnkey, no-
hassle solution, it's probably a good fit. I'm anxious to see how this product
evolves over time. I think it has the potential to become an important
technology in the AS/400 industry as it matures in the years ahead.
--Robin Klima
Tom's introductory question is "FSIOP--Is it the solution?" I don't mean to put
words in his mouth, but I suspect he really means, FSIOP--Is it a solution? I
don't think he is arguing that everyone who has an AS/400 should rush out and
buy an FSIOP. I do think that, in certain very specific situations, the FSIOP
does have a place.
The FSIOP seems to be fit for small shops that have only an AS/400 and don't
have a LAN in place already. The benefits include the ability to call a single
vendor (IBM) if you have problems with the hardware or software. Tom is right
that, for a small shop, the consistency of maintaining the LAN environment in
conjunction with the AS/400 environment can be helpful. If you are in a
situation where you are the data processing manager, computer operator,
programmer, and data entry person, and now your management says you need to
implement a LAN, the FSIOP may be just what you're looking for.
For larger environments and those sites that already have a LAN, it is
difficult to ignore the issues that Tom raised as concerns about the FSIOP.
Those points directly relate to the value of the FSIOP. The value is the
relationship between the cost of the hardware and software (remember, IBM is
supplying this and doesn't make anything inexpensive) and the benefit you gain
from that package.
If there is no easy and inexpensive upgrade path, that reduces the value. If
you have to train someone used to graphical LAN administration tools to use an
inefficient text interface, that reduces the value. If the FSIOP doesn't run
the Network Operation Systems (NOSs) you're running already, all of these
things reduce the value.
The FSIOP has a place in certain shops in certain situations, but it clearly is
not the solution.
--Jim Hoopes
Kudos to IBM for an innovative idea that could prove to be a major contributor
to the continued success of the AS/400. What a great ideaùto consolidate,
integrate, and simplify the myriad of disparate hardware and software
components involved with computer networks.
I think the solution the FSIOP provides is the important thing to look at, not
how the engineers decided to implement it. Look at the benefits the FSIOP
provides:
o It integrates network management, so it's possible for the AS/400
adminstrator to manage the resources of both the AS/400 and the LAN from any
AS/400 terminal or attached PC.
o It adds the AS/400's reliability factor to a LAN.
o Your AS/400 and LAN save/restore strategy can be combined into one integrated
procedure that takes place at one physical locationùthe AS/400.
Hey, this solution sounds pretty good! I'm not saying that you should go out
and replace an existing PC server with an FSIOP. After all, you wouldn't want
to replace your PC LAN administrator; your company might have to pay higher
unemployment taxes.
Sure, there are more powerful chips than the current FSIOP 486/66, but is the
performance of a server simply a factor of CPU power? Isn't file I/O a
significant factor? I'd guess that a 486/66 CPU is plenty of CPU power for
typical network loads. Also, don't forget, you can now install up to four
FSIOPs on a single AS/400. Kris, I wouldn't get too hung up on the CPU.
In a world where computer networks are spreading like wildfire, I think it
makes a lot of sense to give the AS/400 a chance not only to participate, but
to offer a superior solution.
--Richard Shaler
LATEST COMMENTS
MC Press Online