IMHO, "mainframes are going away", at least as we have known them.
Both Steve and Timothy (plus many others) make some very interesting
points. Corporations seem to be overlooking the fact that many of us
older IT workers (I first started on the S/360 when it was fairly new)
will retire in the next few years. I think many of us may have retired
sooner, but for our portfolios be hit over the past few years or our
retirement plans simply being cut.
Like Steve, I currently do training and consulting for a living and am
looking forward to the renaissance in mainframe training and consulting
in the near future. I am not sure that the retiring people of my era
will go into training, but may do a little part time consulting after
they "hit the beach" at least once or twice.
For a few observations I have seen in my travels which may provide some
thoughts for people to consider or at least ponder a little. First our
industry is in the midst of major consolidations. This is being driven
by mostly cost. Companies are reducing the number of data centers to
save on the software cost , as stated in the thread. With the larger,
faster mainframes this is now practical for many sites.
Many corporations are going through mergers, acquisitions, buyouts or
simply reductions in size. This leads to less mainframes, plus less
people to maintain them. One reason our large corporations have no
seemingly urgent need to train people.
Corporations want to "web enable" all of their applications. This means
new interfaces to existing legacy (heritage) applications running on our
mainframes. While most of the new development is aimed at the Linux,
UNIX, Windows (LUW) platform, some changes will be required to the
mainframe applications to handle the removal of the user interface.
This seems to be generation the need for people with multiple platform
skills. We have had some Java people in Cobol classes and Cobol people
in Java classes.
I believe that corporations will not totally eliminate their
mainframes. They have too much invested in the core applications
driving their business. Yes some smaller sites (as mentioned in this
thread) will move to LUW platforms, but those with large databases and
high volume transaction processing could not afford to run their
business on the smaller platforms. The mainframe will realize its
future as the back-end database, transaction processing system for these
large corporations. Many of the recent IBM hardware/software release
seem to be positioning the mainframe into this mold.
IBM is committing to mainframe education. They have started an
initiative with hundreds of professors and universities to bring
mainframe training back to the college world. I am partially saddened
by this (so is Steve I think), since it does take work from people like
me. I also think that there may be enough training to spread around a
little since many people currently in the IT field will want to obtain
training in a short period of time rather then across many weeks. The
universities seem to want you to take a Cobol course in 12 or more weeks
which organizations like mine provide in 2-3 weeks.
Another point to think about, IBM has stated in sessions (like Share)
that many of the smaller mainframes (z800 and z890) do not even run
z/OS. They are running new workloads like Linux and UNIX, either native
or under z/VM. It is still a mainframe, just not the traditional
operating system we know. No DD's in Linux as far as I know.
Corporations want the reliability of the mainframe but with support for
the newer application server environments.
Speaking of WAS, I question if many sites will or are running WAS under
z/OS. I have worked with WAS V5 and V6 under z/OS and let me tell you
that IBM has finally figured out how to create their version of a slow
system. We do not have one of the new fangled zAAPs to assist the Java
work load, but I have even been at sites which do have them and WAS
still is relatively slow due to the fact that z/OS is multi-lingual
(EBCDIC, ASCII and UNICODE). I believe that this is why most of the
application servers seem to be on the LUW platforms since corporations
can tune the work load in support of that world. We as mainframes must
maintain a blend of work loads at some optimum level.
Now the pitch. I for one have been an "old dog learning new tricks" and
do not plan to retire in the near future. I (Steve and others) will be
more then happy to provide training in the mainframe disciplines. We
are having a lot of fun blending the old with the new. I enjoy teaching
Java to my fellow Cobol programmers, and even Cobol to my fellow Java
programmers.
If you have many years invested into a company and its business, then
that knowledge is hopefully invaluable to the company. Take the new web
environments as a challenge to leverage you knowledge into a better way
of presenting information to your users. They will still require your
skills in SQL, plus all of the other legacy (heritage) forms of
extracting corporate information to drive the business.
BTW, this is a forum for mainframe people, plus other interested
parties, but have you had a look at the new i5? The AS/400 has also
learned new tricks including LPARs. It can run many different operating
systems in LPARs all on the same physical piece of hardware. Just one
more way corporations can consolidate their many servers to reduce the
cost of software (which may be of more concern then the hardware in
todays data center). Many of the mainframe technologies developed over
the years now seem to be surfacing on the smaller platforms.
I look forward to being able to apply my knowledge to many different
platforms, just in slightly different ways. I hope many more will seize
the opportunity to investigate this "marriage of skills" and have some
fun while doing it. Never stop the learning process.
Steve Comstock wrote:
Timothy Sipples wrote:
I hope you've been reading this thread because this
is the real world I encounter. IBM seems to have no
sense of urgency over this, but I'm might close to
closing my doors after 30 years of being a self-employed
trainer and 7 years with IBM before that.
My sense of urgency is measured in weeks, not years.
And most companies will try to train in-house to one degree or
another.
The last statement was true, once upon a time. Now, it seems,
very few companies do any training, except for "soft skills".
I would imagine there's also going to be an increasing emphasis on
new mainframe technologies as interest continues to grow: WebSphere
Application Server, Web services, Linux, Java, etc. Which is not to
say that everything else is going away -- far from it -- but the
ratio could change.
Sometimes ya' simply guess wrong. I put my energies into
developing courses for supporting the Web on the mainframe:
z/OS UNIX (3 courses)
X/HTML on z/OS
To be announced next week: WebSphere Developer for zSeries
But the prospects I talk to are not moving that direction.
And the IBM sales reps aren't talking that direction.
I'm not exactly sure yet how IBM retirements will affect the supply
of mainframe trainers and, thus, the training market. There are too
many cross currents to get a good read. I believe IBM has announced
that it's in a mainframe hiring mode, and that's been my personal
observation, too.
Need any training?
Re: IBM and a sense of urgency, the big reason I'm where I'm at this
moment is because of a corporate sense of urgency. So, at least in
my personal experience (and observation), yeah, we got that -- so
much so that I'm roughly halfway around the world from home.
--
Regards,
Thomas Dunlap Chief Technology Officer [EMAIL PROTECTED]
Themis, Inc. http://www.themisinc.com 1 (800) 756-3000
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
