Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations Chriss Miller on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
Where does the Mainframe fit in?
- Thread starter mrobbo
- Start date
I would suggest that your question relates to both MainFrame and UNIX skills.
It seems to me that, over the next five years or so, we will be seeing more proprietary Internet and fewer Open Systems (or Open Source for that matter) solutions.
There.... That should get me flamed... [sig]<p>Mike<br><a href=mailto:michael.j.lacey@ntlworld.com>michael.j.lacey@ntlworld.com</a><br><a href= Cargill's Corporate Web Site</a><br>Making mistakes, so you don't have to. <grin>[/sig]
It seems to me that, over the next five years or so, we will be seeing more proprietary Internet and fewer Open Systems (or Open Source for that matter) solutions.
There.... That should get me flamed... [sig]<p>Mike<br><a href=mailto:michael.j.lacey@ntlworld.com>michael.j.lacey@ntlworld.com</a><br><a href= Cargill's Corporate Web Site</a><br>Making mistakes, so you don't have to. <grin>[/sig]
I actually learned and grew up on the mainframes but for 23 years now have been working with mid-range and for the last 10 unix servers. The speed, flexibility, robustness, cost, etc,etc,etc of the unix seems to address every question any of my clients have had and then some. I believe that the large legacy companies who have invested millions into mainframe development are keeping these systems so experience return from that huge investment. Perhaps (and this is already being done) is to allow the mainframe OS's and apps to run (in emulation) on the unix boxes. IBM already does with their OS/390 emulator on to op AIX.
So, bottom line from my point of view, is that why should any NEW design or development be performed for the mainframes?
So, bottom line from my point of view, is that why should any NEW design or development be performed for the mainframes?
Everyone here knows how fast technology changes and i think very few computer people will be willing to buy a hugh mainframe to support their webapplications. Most people would go with a boosted IBM PC or maybe a sun microsystems computer. With these system you can still run unix and linux for web applications.
I think the big mainframes will still have a place but my guess is only for huge databases and storage backup stuff, perhaps...
[sig]<p>ackka<br><a href=mailto:tmoses@iname.com>tmoses@iname.com</a><br><a href= my site</a><br>"Do No Harm, Leave No Tracks"<br>
ICMP Summer 2000, 2600 Article<br>
<br>
[/sig]
I think the big mainframes will still have a place but my guess is only for huge databases and storage backup stuff, perhaps...
[sig]<p>ackka<br><a href=mailto:tmoses@iname.com>tmoses@iname.com</a><br><a href= my site</a><br>"Do No Harm, Leave No Tracks"<br>
ICMP Summer 2000, 2600 Article<br>
<br>
[/sig]
COBOL
MIS
- Jul 2, 2000
- 26
Also I like many are asking when shall the Computer Biz
Stabalize...? I also have a lot of MF and Unix Experience.
A few months ago at a IBM AS400 Roadshow we were told that in contrast to Linux Very few people have Access to the
Operating System Source Code of OS/400!
This Makes AS400 one of the most safe Systems available...
Keeps the Hackers OUT!
AS400 Systems are very Scalable and Easy to Upgrade.
The Software just goes on running for years...
[sig][/sig]
Stabalize...? I also have a lot of MF and Unix Experience.
A few months ago at a IBM AS400 Roadshow we were told that in contrast to Linux Very few people have Access to the
Operating System Source Code of OS/400!
This Makes AS400 one of the most safe Systems available...
Keeps the Hackers OUT!
AS400 Systems are very Scalable and Easy to Upgrade.
The Software just goes on running for years...
[sig][/sig]
COBOL,
Don't think I'm starting an AS/400 vs Linux debate here, and if one starts I'll get bored with the thread.
"AS400 Systems are very Scalable and Easy to Upgrade" -- yes
"The Software just goes on running for years..." -- yes
Just picking up what you said about the number of ppl with access to OS/400 source. I'm not aware of many instances of Linux being hacked -- that is of someone inserting something nasty into the source code and that getting into a standard distribution.
It could be argued that the more pairs of eyes that can look at the source then the higher the chance of something nasty being picked up... [sig][/sig]
Don't think I'm starting an AS/400 vs Linux debate here, and if one starts I'll get bored with the thread.
"AS400 Systems are very Scalable and Easy to Upgrade" -- yes
"The Software just goes on running for years..." -- yes
Just picking up what you said about the number of ppl with access to OS/400 source. I'm not aware of many instances of Linux being hacked -- that is of someone inserting something nasty into the source code and that getting into a standard distribution.
It could be argued that the more pairs of eyes that can look at the source then the higher the chance of something nasty being picked up... [sig][/sig]
I really have to agree with MikeLacey on this one,
<rant>
for example the windows9x source has never been revealed, but as you know a great deal of bugs, packet floods etc.. has been released to the genereal populace. Chances are if everbody used OS/400 just as many bugs would have appeared. Besides the native assembly language code is still readable to anyone with a large amount of time.....It is really not security.....
If you want to continue this discussion we should probably move to genereal secuirty desussion under ms/it stuff
later [sig]<p>ackka<br><a href=mailto:tmoses@iname.com>tmoses@iname.com</a><br><a href= my site</a><br>"Do No Harm, Leave No Tracks"<br>
ICMP Summer 2000, 2600 Article<br>
<br>
[/sig]
<rant>
for example the windows9x source has never been revealed, but as you know a great deal of bugs, packet floods etc.. has been released to the genereal populace. Chances are if everbody used OS/400 just as many bugs would have appeared. Besides the native assembly language code is still readable to anyone with a large amount of time.....It is really not security.....
If you want to continue this discussion we should probably move to genereal secuirty desussion under ms/it stuff
later [sig]<p>ackka<br><a href=mailto:tmoses@iname.com>tmoses@iname.com</a><br><a href= my site</a><br>"Do No Harm, Leave No Tracks"<br>
ICMP Summer 2000, 2600 Article<br>
<br>
[/sig]
Getting back to the original question... " but how relevant is this likely to be in the future?" In my opinion, systems running with COBOL and DB2 will still be quite relevant because there is so much COBOL code in use throughout the world it cannot be easily or cheaply replaced by other languages. It simply isn't feasible or cost effective, so these apps will continue to run. Those of us "dinosaurs" will have plenty of maintenance to do if nothing else. With regard to UNIX, you're going to see more and more Linux running under OS/390 and the new z/OS -- and it requires NO emulation. Both operating systems can run many, many Linux images simultaneously and they are extremely secure. When was the last time you heard of someone hacking into a "mainframe" OS? Linux will also run on AIX boxes so you'll have some growth there also. I think someone mentioned something about big expensive mainframes in one of the responses... they ain't so big these days. We run OS/390 on IBM's Multiprise 3000 and this machine is way more powerful than the 3090 we had 4 years ago, and it is smaller than a dishwasher. Not to mention the uptime factor... our mainframe NEVER has gone down -- can't say that about an NT box. COBOL is old for sure, but it will be around for a long time to come. I think the skill to add to your COBOL and DB2 will be Java which also can run under CICS on the IBM boxes. If things go according to plan, our web site will be running under Linux on the "mainframe" and we'll kiss the old site's IIS/NT box good bye. A lot has happened during my 35 years in the business and I can hardly wait to see what's next.
Jim
the mainframe dead? -- no way!
Jim
the mainframe dead? -- no way!
I'll just add my 2 cents and agree with cobol. Though I'm a VB developer, my firm is primarily an AS/400 shop, has been from day 1, and S/36 before that. Over the years, our NT boxes go down weekly (no surprise), but I've seen our hp9000 go down, and even our rs/6000/aix fail. But we have 2 AS/400's, both used for heavy development, training and our internal business systems.
We have brand new programmers who don't know what they're doing trying all sorts of strange, bad code on them, and highly experienced cutting-edge rpg (can rpg be cutting edge?) programmers testing the limits of these machines with everything they can throw at them, and every kind of user in between. I've been at this firm 4 years and neither one has *ever* abended. Ever. Nor have they been hacked, and they're on the internet with clients dialing constantly. I'm primarily a PC developer, but I've gained a true respect for the 400.
I now have a project using VB over DB2 that's about to go to a client, and not only is the performance absolutely blistering--the most amazing thing is that when we install the system at the client, I can truly sleep at night knowing that no matter what kind of maniacal stuff the users will do, the machine will not go down. After so many NT based apps--that's a strange and wonderful feeling.
--Jim
We have brand new programmers who don't know what they're doing trying all sorts of strange, bad code on them, and highly experienced cutting-edge rpg (can rpg be cutting edge?) programmers testing the limits of these machines with everything they can throw at them, and every kind of user in between. I've been at this firm 4 years and neither one has *ever* abended. Ever. Nor have they been hacked, and they're on the internet with clients dialing constantly. I'm primarily a PC developer, but I've gained a true respect for the 400.
I now have a project using VB over DB2 that's about to go to a client, and not only is the performance absolutely blistering--the most amazing thing is that when we install the system at the client, I can truly sleep at night knowing that no matter what kind of maniacal stuff the users will do, the machine will not go down. After so many NT based apps--that's a strange and wonderful feeling.
--Jim
Guest_imported
New member
- Jan 1, 1970
- 0
Unfortunately, the trend seems to be taking us down the same path that it's been on for the past few years. PC LANs that still use the same basic underlying technology that the old 8080 procesors used. Sure, chips have gotten faster, disks bigger/cheaper and network connections faster but it's still running on a archaic IRQ, I/O Bus, hardware driver based system. The software situation is getting even worse. Proprietary languages that you hope never go out of style. Hardware drivers that cease being supported forcing hardware upgrades. Having to buy bits and pieces of software and then hope they all work together (sure the sale rep says they will but c'mon we all know it's not always true).
Maybe it's just me but I like the idea of bringing in a mainframe. I can have it up and running in about two hours, run code 20 years old if I need to, use printers and other devices that are 15 years old or older, have an uptime percentage of 99.9999999 (24 X 7 X 365 X 20 is a reality), get subsecond response time, barring a cataclysmic disaster never lose one byte of data, use it as a server that will blow away any PC server by a thousand fold and the best thing, do it cheaper than a PC based LAN all on a refrigerator sized box sitting in a corner.
Just IMHO.
Maybe it's just me but I like the idea of bringing in a mainframe. I can have it up and running in about two hours, run code 20 years old if I need to, use printers and other devices that are 15 years old or older, have an uptime percentage of 99.9999999 (24 X 7 X 365 X 20 is a reality), get subsecond response time, barring a cataclysmic disaster never lose one byte of data, use it as a server that will blow away any PC server by a thousand fold and the best thing, do it cheaper than a PC based LAN all on a refrigerator sized box sitting in a corner.
Just IMHO.
doublehelix
IS-IT--Management
At my workplace I suggested we deep freeze all our COBOL programmers and thaw 'em one at a time when we needed to perform maintenance on our clunky old legacy systems. And I wonder why people don't take all my recommendations seriously...
dilettante
MIS
The only real reason most people still run the older types of systems is that they have a lot of important code there.
It just isn't quick or easy to port all of that stuff at the same time you have to keep updating it.
Reliability is becoming less of an issue. Most mainframes go down when they do now not because of hardware, or even software, but because most of the talented people who operated and administered them have retired or moved on to greener pastures. It just wasn't viable to stay in the mainframe arena, where pay scales are narrow.
That doesn't mean there isn't good money there, just that the ceiling is so much lower than in newer technologies.
As a result many mainframes are now administered by ex hardware jockeys with little formal background and a few tech classes, and often operated by part-timers and students who don't even know the system. These people are struggling - doing their best - to keep things going, but it is amazing how we've gone from 99.5% uptime (and better) to the problems we have today due to the lack of appreciation of knowledge and ability. Anybody seeing even 99% uptime from a mainframe today must really have good people yet.
But things are even scarier in the mini/micro arena! Hardware is less reliable (but getting better all the time). The "people story" is if anything even worse though!
I've never seen such a fertile ground for nepotism and getting by without credentials than in the "network administration" arena. The answer to everything seems to be "reboot and hold your breath." And no, I don't think "certification" style credentials are any answer. I work with plenty of paper-cert types who can't tie their shoes.
Cobol needs to die, but can't. The biggest problem is that there just isn't a good successor language yet - and the answer isn't yet another C variation. Structs frankly are poor as a metaphor for records, and this is an area where Cobol still excels, despite some nasty primitive syntax.
Ada should have been "it" - but clearly Ada's architects tried to bite off too much just like IBM did back in the 60s with PL/I. Nobody uses Ada who has a choice about it.
Oh well, this is just depressing me. ;-)
It just isn't quick or easy to port all of that stuff at the same time you have to keep updating it.
Reliability is becoming less of an issue. Most mainframes go down when they do now not because of hardware, or even software, but because most of the talented people who operated and administered them have retired or moved on to greener pastures. It just wasn't viable to stay in the mainframe arena, where pay scales are narrow.
That doesn't mean there isn't good money there, just that the ceiling is so much lower than in newer technologies.
As a result many mainframes are now administered by ex hardware jockeys with little formal background and a few tech classes, and often operated by part-timers and students who don't even know the system. These people are struggling - doing their best - to keep things going, but it is amazing how we've gone from 99.5% uptime (and better) to the problems we have today due to the lack of appreciation of knowledge and ability. Anybody seeing even 99% uptime from a mainframe today must really have good people yet.
But things are even scarier in the mini/micro arena! Hardware is less reliable (but getting better all the time). The "people story" is if anything even worse though!
I've never seen such a fertile ground for nepotism and getting by without credentials than in the "network administration" arena. The answer to everything seems to be "reboot and hold your breath." And no, I don't think "certification" style credentials are any answer. I work with plenty of paper-cert types who can't tie their shoes.
Cobol needs to die, but can't. The biggest problem is that there just isn't a good successor language yet - and the answer isn't yet another C variation. Structs frankly are poor as a metaphor for records, and this is an area where Cobol still excels, despite some nasty primitive syntax.
Ada should have been "it" - but clearly Ada's architects tried to bite off too much just like IBM did back in the 60s with PL/I. Nobody uses Ada who has a choice about it.
Oh well, this is just depressing me. ;-)
- Status
Similar threads
- Locked
- Question