No Coder Left Behind
A Mainframe easy enough for my brother's kids
As escalating IT security threats, datacenter utility bills and other market conditions continue to drive new customers to the mainframe, we're noticing an increased demand for mainframe skills. While a cadre of initiatives over the past few years like the Master the Mainframe contest have produced over 15,000 new mainframe skills, today IBM announced a goal to turn every computer science student in the world into a mainframe-capable coder or administrator by 2011. This is supported by a $100 million investment to make the platform so simple that even a windows programmer can use it <g>
As the character in Mainframe meets "The Office" said, "let me explain it to you like you're my brothers' kids:" The new z/OS V1.8 includes a sharp GUI, which stands for "pretty pictures," as well as other tools to make life simple, including a "personal trainer" which checks out the system and makes adjustments and fixes. And a new web resource, the z/OS Basic Skills Information Center, designed specifically for IT professionals who are new to the mainframe and z/OS.
Coders, switch to the mainframe. No skills?... no problem.
|by rjhoey||October 4, 2006 in Application Development |
TrackBack URL for this entry:
Listed below are links to weblogs that reference No Coder Left Behind :
» Extreme Makeover: z/OS Edition from Getting the Most out of DB2 for z/OS and System z
Question: If you were 16 years old living in 2006, would you be more interested in a Mac or a z9? While you're trying to decide, is there something we can do to help influence your decision? Is there anything... [Read More]
Tracked on Oct 6, 2006 2:01:40 PM
25 years ago, I almost did not cross over into systems programming because someone said "all systems will be turnkey by the end of this decade". Fortunately for me, I ignored that prediction.
While I applaud all efforts to make it nice and simple, I have this image from "The Wizard of Oz" where we system programmers will still be playing the role of the wizard, albeit from behind the curtain as usual.
Posted by: Bob Richards | Oct 5, 2006 7:15:19 AM
No doubt, big systems will always need wizards. This simplification effort is about masking complexity for the administrator and developer wherever possible – for a certain % of the roles. The objective is to lower the bar for entry, and to simplify some of the mainframe technical career paths to allow more folks to get involved on the platform.
Deeper skills will be required for areas like Enterprise system architecture and disaster recovery planning, but we want to dramatically reduce the need for those skills where possible.
Thanks again for the comment.
Posted by: Bob Hoey | Oct 5, 2006 6:53:47 PM
I wonder what percentage of CS students IBM management thinks will be interested in becoming a mainframe coder/admin, even if the CS students figure they have the ability to do so....
Posted by: Jonathan Edwards | Oct 6, 2006 11:08:12 AM
Admin? That is a *nix term.
It does not matter how many they think want to become "z/OS systems programmers"... it matters how many actually do!
Posted by: Bob Richards | Oct 7, 2006 9:43:14 AM
>>>I wonder what percentage of CS students IBM management thinks will be interested in becoming a mainframe coder/admin, even if the CS students figure they have the ability to do so....<<<
Anyone who learns WebSphere development is automatically a "mainframe coder," since any WAS app can be deployed to the mainframe. And I think that once students discover how much more attractive their CV/resumes are to prospective employers when they see "z/OS" skills listed, a significant percentage of them will get interested quickly...
Posted by: Bill Seubert | Oct 10, 2006 10:54:30 PM
Simply because there is going to be a nice GUI does not mean companies will start allowing everyone to play with them. Employers will still looks for skills when hiring someone to run their mainframes. The only way to increase the pool of mainframe profesionals is to increase the availablity of systems where interested persons can play and learn. One way to do it would be to have a hobbyist license (which is cheap enough for the masses) to run z/OS under some emulator.
Posted by: M. Khalid Khan | Oct 16, 2006 3:23:43 PM
I have just started doing an internship programme with IBM. I am doing mainframe(Z/OS). I always enjoy reading news about whats happening at IBM. It feels so courageous to find that Mainframe skill is on demand at the moment and in future.This makes me enjoy and have a will to learn more as a new IT graduate.
Posted by: Mokaba Ivy | Oct 31, 2006 6:48:41 AM
Khalid wrote: "The only way to increase the pool of mainframe profesionals is to increase the availablity of systems where interested persons can play and learn. One way to do it would be to have a hobbyist license (which is cheap enough for the masses) to run z/OS under some emulator."
I only agree with this to a point. I think more accessibility to the OS and software would improve things, but putting a copy of z/OS in the hands of an untrained person is a little risky. Inside IBM, I have seen numerous examples of untrained technical folks getting copies of z/OS to "play with," and they inevitibly get frustrated and give up, and things are worse off than they were to start with.
The problem must be addressed from several angles. The announcement of the investments in System z usability indicates that we are tackling the problem from one direction - reducing the skill requirements for the mainframe. Khalid, your proposal is one approach to attacking the problem from the other direction - building skill. IBM has programs to address the "skill gap" from both directions (reduce needed skill, improve skill availability), even though there is no "program" right now (that I am aware of) to put copies of z/OS or other zSW in the hands of "hobbyists".
Posted by: Bill Seubert | Nov 7, 2006 2:42:44 PM
Geez, someone has to spend $100M to entice fresh talent to spend time with a mainframe? Pathetic.
The reason why there is a GUI is because the "system programmers" are too old, have failing eyes and need some nice big pointers and boxes to click in.
PS I've never felt something push so much heat as a mainframe. The massive amount of electricity that it sucks is being wasted in the form of heat. Rediculously overpriced when compared to a bladecenter.
Posted by: Penguin | Nov 10, 2006 12:04:52 AM
>>>PS I've never felt something push so much heat as a mainframe. The massive amount of electricity that it sucks is being wasted in the form of heat. Rediculously overpriced when compared to a bladecenter.<<<
Penguin, you might wish to do some research on power consumption for server farms. While a single bladecenter may consume less power and generate less heat, seldom can one bladecenter do the same amount of work as a single mainframe. It usually takes hundreds of servers to equal the workload capacity of a mainframe. The average utilization of an Intel server runs somewhere between 5-10%. The average utilization of mainframe systems runs around 60%, and customers routinely run their production systems at or near 100% for long (hours and hours) intervals, with no impact on workload performance. As a result, it is not at all uncommon to see ratios of 15-30 distributed engines to one mainframe engine, or even higher. Those blades sit there churning out heat and consuming power, all the while doing little or nothing.
A recent Wall St. Journal article noted that the typical distributed server farm now consumes an average of 3200 watts per sq. foot of floorspace. The average mainframe consumes only around 400W/ft2. A recent study showed that replacing a 1000 server farm with a mainframe would save between $400K-$600K in power and cooling alone, based upon average kWH utility prices and power consumption of mainframes vs. distributed servers.
And regarding "overpriced," most exhaustive total cost of ownership surveys show that the cost of ownership of mainframe systems is less than equivalent distributed implementations, when cost of HW/SW acquisition, environmentals, and the cost of administration are taken into account. The cost of people is now around 50% of the cost to run a computer system, and the mainframe runs rings around distributed in that area. And, even on the "total cost of acquisition" front, acquiring a mainframe (HW and SW) can be cheaper than a server farm, since we now have very inexpensive low-end (but high capability) mainframes with the System z9 BC.
Posted by: Bill Seubert | Nov 10, 2006 1:55:55 PM
The comments to this entry are closed.