Re: Ibm i modernization

So Dr Franken knows his stuff and I second almost of what he says and other folks on this thread.  I program in multiple languages on IBM i and other platforms, including the cloud, and it comes down to what Joe Pluta often said in these forums: Use the best tool for the job.  You want snappy front end stuff, go with Node and frameworks like Angular or whatever flavor of the day suits you.  But, for backend stability, security and reliability, I find that RPG and Java serve me just fine, thank you.  Performant, mature and stable.

Two additional thoughts:

1)  12 years ago I took a secure programming class and there were two programmers from Costco in the class.  I asked them how they were going to utilize the class in an iSeries environment.  They said they weren't.  That Costco was moving off of iSeries and their job was to make sure the new application was as secure and stable as what they had with the "old green screen apps on the AS/400".  How did that turn out?  You Costco members  know:  Green screen rules the day for fast, secure, stable and efficient access to data at Costco.

2) Buy yourself a copy of "Kill It with Fire: Manage Aging Computer Systems (and Future Proof Modern Ones)" by Marianne Bellotti.  It's a reflection on what she learned in modernization on the mainframe.  GREAT book.  It shows the right way to go about modernization and "Killing it with Fire" is NOT the first step....

Pete Helgren
www.petesworkshop.com
GIAC Secure Software Programmer-Java
GIAC Cloud Penetration Tester
AWS Certified Cloud Practitioner
Microsoft Certified: Azure Fundamentals

On 6/17/2023 4:11 PM, Larry "DrFranken" Bolhuis wrote:

In a presentation from IBM recently the idea that IBM i is the 'old legacy server' was completely destroyed.  Steve Will, IBM i's chief architect, and an IBM Distinguished Engineer, along side of Bill Starke, who is the Chief Architect for the POWER Processors, also a Distinguished Engineer covered the history together.

What's old? Mainframe! Younger than that are the Unixes including AIX and Linux.  Then you get into the chain that produced DOS and Windows. Finally the 'new kid on the block' was the System/38 and a decade later the AS/400, now IBM i.  So this whole business of 'we're old' is actually backwards. We're the youngest at the table. We STARTED with Single level storage and so much more.

Then consider the hardware on which we run. Most reliable servers for the past decade in a row now. Most scalable as well. From the smallest single core to the largest E1080 the same software will run and run happily even if the processors are north of 90% busy. And the systems support vast amounts of RAM and more than double the memory throughput of any Intel system.

And of course we have Db2. Not just a good database, a great one! And it keeps getting better every day. And it can be mirrored to another system with Db2 Mirror, setting our sites on true continuous uptime.

And let's don't forget all the other hardware bits that are supported like NVMe for example, as well as PCIe gen 5 slots, and new forms of memory that are more reliable and faster. Then connect the things with network of 100Gbps and you have one serious workhorse there!

So just drop any pretense of moving to another platform for performance, reliability, scalability, or any new hardware bits. Any attempt for that is misguided at best.

As has been mentioned, breaking up your application into many layers on different platforms some in cloud, some in the data center, some on Linux, some in Windows, and on and on, means you have SO MANY moving parts that something is bound to break at any second. And a simple change in one area could cause havoc in another. Lose a few folks to job changes and it's probable that NOBODY can put humpty dumpty together again. Planning a true HA scenario is also fabulously more complicated.

Now your applications? Yeah it's true that IBM i can actually run programs from S/34, S/38 and S/36, and some of you do (you know who you are.) And that may be perfectly fine as they do so correctly and properly support your business. However, many others are feeding those (ahem) 'way cool' subfiles on which you can page forever. Said subfiles fail miserably on the basic tablet, smart-phone, etc.  But the database and the logic around that, while almost CERTAINLY not perfect, contain an untold wealth of knowledge. Transferring that knowledge into a new language or database, and getting it right and complete, will not take months, it will take YEARS.  The number of such projects that have completed to great success that I am aware of can be counted on my ring finger.

Do 'off the i' projects complete? Sure they do. Some complete with the company going out of business. Some complete with the company being sold. Some complete with the company just declaring success, and dealing with the mess they have made.

What's  INCREDIBLE to me is how often the new app is asked to do things the old app did for decades. Sure, sometimes it's the old 'we've always done it that way' problem, but often the issue is a key process to the business.

Here is a classic case. Company that sold Conveyor belting (along with all the supporting bits) had a powerful application to determine the minimum inventory levels needed at each location. The process understood that a 100 by 8 ft piece of belt could be cut in both the long and short directions. Doing so leaves two additional pieces each of which could fulfill another minimum. When they moved off IBM i, the new software's solution was 'use square inches.' Just 'do the math' and assure you have enough sqins of each kind of belt. How hard is this?  It was a disaster of course and the company is no longer in business.

IBM i has so very many strong capabilities and supports so much that there is 'always a way' to accomplish a needed task.

As has been mentioned it's the user interface where most of these applications lag. And that CAN be done by non IBM i developers. Look at what Jack Woehr said in this thread.  Use the skills of each type of developer to turn your IBM i application into a truly modern application that still remembers all those special considerations built into the application to make your business the best it can be.

    - Larry "DrFranken" Bolhuis
    IBM Champion

On 6/17/2023 2:36 PM, Jack Woehr via MIDRANGE-L wrote:

On Sat, Jun 17, 2023 at 11:53 AM Efield <netnature212@xxxxxxxxx> wrote:

1) So why does everyone wants to go after this modernization for Ibm i?

Because it's the only way to deliver the sorts of web apps and dashboards
an enterprise user base and leadership expect from a system these days, not
to mention externals like customers.
It's straightforward enough a process to modernize without leaving the IBM
i platform if you have the folks who are motivated and ready to do the work
involved.

2) Why is Ibm i professionals are so hard to find and why are so less
opportunities for them and why are there salaries no where match to full
stack developers

The second clause of your question answers the first clause!

,why is it to hard freelancing jobs for IBM I compared to
other open source technologies like python, java, cloud etc. ?

See your previous question.

3) Why is everyone running after cloud computing platform?

It's tempting to distribute your apps on serverless hosting, but the cost
and reliability comes back to bite you.
The reliability and security of a single unified platform properly
disaster-recovery-ready is more to the advantage of the enterprise.
Of course you can cloud your IBM i LPARs which is very convenient,
reliable, and disaster-recovery-ready. Absolute Performance, my company,
has a Power cloud.

4) Why should users of IBM I applications have right to get mobile

app/weblinks kind of easyness in terms of their day to day operations?

That's a good idea, and there are many solutions for that.

5) Any idea how many apps based on IBM I around the world have been
decommisioned from this IBM I platform and have been modernized now?

  Modernization does not require leaving the IBM i platform!