 |
|
Wobbling WebSphere — A Special Report
Will WebSphere Express solve the low-end iSeries conundrum?
by Joe Pluta
The complete discussion can be found here.
No quotes today. No cute comments. This special report is
about a very serious issue:
the future of the WebSphere Application Server (WAS).
Some of you may remember back to last year when IBM first
announced WAS Version 4. I was pretty vocal about it because IBM discontinued
the free Standard Edition of WAS (WAS 3.5 SE), instead offering only the rather
pricey Advanced Edition (currently selling for about $8,000 to $12,000 per CPU).
I even used the term "bait-and-switch" because IBM had specifically directed
us--both developers and end users--to use WebSphere as a development platform.
Those of us who listened got a rather rude awakening when we found that we would
have to add a significant premium to our software costs in order to continue to
use something that IBM had originally given us for free, and we made our
position known.
In response, IBM announced the support of Tomcat, the
open-source Web application server. And while that's been an interesting option,
we were never sure whether it was an interim solution, a strategic direction, or
something in between. And the fact that IBM doesn't ship the latest version of
Tomcat (IBM ships the older Version 3 instead of the much faster and more
feature-rich Version 4) doesn't help clarify the issue.
The release of
V5R2 of OS/400 muddies the waters even further. This release no longer supports
WAS 3.5 SE. In fact, this release breaks it. That's right: If you are using WAS
3.5 SE and you upgrade to V5R2, you are dead in the water. You will be forced to
move to a different application server, either Tomcat or WAS Advanced Edition.
No matter how IBM chooses to spin it, this is particularly foul treatment of
both developers and users.
However, some good news is in the offing. At
COMMON, you'll hear about WebSphere Express, the "low-cost" alternative to WAS
Advanced Edition. Sources tell me that you'll hear an official announcement in
Q4, and the product will probably be released in Q1 of 2003. But WebSphere
Express isn't for everyone. IBM has firmed up its policy: If you have a machine
with under about 500 CPW, you should run Tomcat. Tomcat will continue to be
shipped with OS/400 and is the Web server of choice for "smaller footprint"
machines.
As to pricing for WebSphere Express, how low is low? Good
question. Even at half the price of the Advanced Edition, you're still talking
thousands of dollars--in effect, a "WebSphere tax"--just to move to V5R2. Note
that this added cost also applies to the "free" WebFacing product, since
WebFacing requires WebSphere. The justification is that the iSeries is the only
platform that has a free version of WebSphere. That may be, but it seems to me
that the iSeries also has a much higher hardware markup per CPU cycle than most
of the other boxes, and those costs have traditionally gone to offset the price
of the software. This is the concept of an integrated machine.
On its surface, this strategic position seems to make some sense. In effect, IBM is
saying that it no longer supports WebSphere for low-end iSeries machines. There
is an implied corollary that IBM won't be making small machines, that the
smallest iSeries machines will soon have enough power to run WebSphere Express.
It's too early to tell what this means as far as pricing for the low-end
user.
But what's the reality of the situation?
If you're running
WAS 3.5 SE today, you must switch your software to either Tomcat or WebSphere
Express prior to the move to V5R2. The move to WebSphere Express will cost you
money, probably thousands of dollars.
If you write iSeries software and
you want to both support low-end users and stay strategic with WebSphere, you
will now have to double your development effort and support both Tomcat and
WebSphere.
So what does all this mean to me as a developer? Well, it
means that I have to either support Tomcat or cut out my low-end users.
Personally, I refuse to do the latter. The low-end customers, the ones who can't
easily justify the continuous upgrades to "bigger and better" machines,
especially in these economic times, are the ones who most need the newer
technologies in order to compete. And even if they can justify an upgrade to a
larger machine, why must they also come up with the added cost of WebSphere
Express, just to stay in line with IBM's strategic direction? And what's to say
IBM won't turn around and discontinue WebSphere Express in the next release of
OS/400?
Since Tomcat is supported on all iSeries platforms, it seems that
it may make sense to save the money and go with Tomcat. And once you've made
that decision, you may decide that running Tomcat on the iSeries is a waste of
cycles. You can buy a dedicated Unix machine running something like FreeBSD for under $700 and move your entire
Web application serving workload off the iSeries. Then, you can use a product
like PSC/400 to convert your
interactive programs to batch, begin the move to true client/server computing,
and avoid an upgrade altogether.
IBM could have avoided this. IBM could
have made WebSphere Express a lean, mean Tomcat-killing entry point into the
WebSphere family: a free, limited-functionality replacement for WAS 3.5 SE.
Instead, IBM has fragmented the user community for what looks to me to be purely
short-term revenue. Why short-term? Because this sort of corporate policy is
what makes developers rethink their positions. When I had to choose between
Tomcat and WebSphere, I chose WebSphere, touting its capabilities to everyone
who would listen. IBM is rewarding that loyalty by now making me choose between
supporting both Tomcat and WebSphere, or supporting Tomcat only. Not only that,
I get to spend extra money if I choose to support WebSphere.
It's becoming an easier decision.
Joe Pluta writes
the "Weaving WebSphere" column for MCMagOnline and is the founder
and chief architect of Pluta Brothers Design, Inc. He has been working in the
field since the late 1970s and has made a career of extending the IBM midrange,
starting back in the days of the IBM System/3. Joe has used WebSphere
extensively, especially as the base for PSC/400, the only product that can move
your legacy systems to the Web using simple green-screen commands. He is also
the author of the popular MC Press book
E-deployment: The Fastest Path to
the Web.
You can reach Joe at joepluta@plutabrothers.com.
Reprinted with permission from MC Mag Online, published by MC Press, LLC;
http://www.mcpressonline.com
|
 |
|
 |
|
|
|
