Debunking Legacy Myths
August 22, 2018 Alex Woodie
In the rush to modernize IT and digitally transform our businesses, legacy systems like IBM i and System z come under increased scrutiny, which is fair. After all, blind loyalty to a particular platform – even Big Blue’s historically stable ones — is a recipe for failure in a dynamic business environment. But when technology providers make questionable claims about the nature of legacy systems in an attempt to sway decision-making, those claims should bear as much scrutiny as the platforms do themselves.
Which brings us to the recent e-book published by First National Technology Solutions and co-sponsored by Dell EMC. Titled “Overcoming the Challenges of Your Legacy Systems,” the 20-page document purports to provide a path forward for those struggling with IBM’s venerable midrange and mainframe platforms.
The document’s authors make some valid statements about nature of legacy platforms in today’s IT environment, and make some verifiable claims that won’t ruffle anybody’s feathers. These legacy platforms do run the core applications that have powered manufacturing, transformation, insurance and finance businesses around the world for decades.
The paper states that legacy systems today need to integrate with “the rest of the systems” to stay relevant, which is a valid point. Many companies are adopting applications that are accessed via individual micro-services that can be tweaked and scaled independently. This is something that can be difficult to do with legacy applications running on monolithic cod stacks, and is a real concern. The shift to DevOps is another the authors point out.
The e-book authors also touch on other sore points with the legacy crowd, such as the difficulty in finding IT professionals with the right skillset to maintain existing applications. Maintaining a population of qualified technicians is always a concern for every platform, but it’s perhaps even more critical for non-mainstream technologies, such as IBM i and the System z mainframe, which have installed bases of perhaps 100,000 and 5,000 (or fewer) organizations, respectively, compared to millions for standards-based X64 systems.
While COBOL and RPG continue to power a huge number of real-world applications on the mainframe and IBM i, it can be challenging to find younger folks with COBOL and RPG skills or the desire to learn “legacy” programming languages like COBOL and RPG. The e-book authors rightly point this out as a problem and a barrier to modernization and scalability.
However, the authors make a few leaps that bring the veracity of the entire e-book into question. For starters, the e-book, which was published earlier this year, makes a trio of unforced errors when it comes to name of the system, the latest version, and the support outlook from IBM.
The e-book stated:
“IBM’s latest version of the AS/400 (IBM iSeries 7.1 introduced in 2010) is still used by the vast majority of users. However, IBM announced that they would be ending support for IBM iSeries 7.1 by the end of April 2018. This puts a spotlight on the end of life/warranty issues that are inherent to legacy technologies with significant CAPEX replacement costs and OPEX maintenance costs.”
Actually, the latest version of the AS/400 is not IBM iSeries 7.1. The name of the operating system is IBM i, and the latest version is 7.3. IBM i 7.1 launched in 2010 and killed it off this April, making it longest-lived operating system release in the entire history of the midrange line, from AS/400’s initial release in 1988 until now.
End-of-life issues are things that IT professionals need to keep in minds, but in the case of IBM i, users will be getting a lot more useful life than practically any other business computer in existence. IBM’s roadmap for the IBM i actually goes out to 2026, and FNTS’ mischaracterization of that fact leads to serious credibility questions.
The authors made another series of gaffes regarding the level of maintenance that legacy platforms require. They stated: “Legacy mainframes require more maintenance time, which leads to higher OPEX as they age and new technology platforms become the norm.” They also stated: “The challenge with legacy systems is that they require constant maintenance and complex updates to keep up with digital transformation needs.”
Every computer ever invented requires maintenance and updates to function well, and the ones that run the core applications of the largest companies on the planet are no exception to that rule. However, FNTS’ insinuation that mainframes and IBM i server require more maintenance than X64 systems runs counter to the available facts.
While one may not believe IBM’s own studies showing IBM i requires a fraction of the support personnel than Windows and Linux systems require, it’s tough to discount the experience of hundreds of IBM i shops over the years, who have lauded the platform for the exceptionally low amount of maintenance it actually requires.
In many respects, the IBM i server’s capability to run at a high utilization rate for long periods of time without undue fiddling and hand-holding by technicians is the single most important reason for choosing the server in the first place. Glossing over this fact with broad characterizations of extensive and expensive personnel requirement compared to X64 does a disservice to those who have spent the bulk of their professional lives working on the IBM i platform.
FNTS gets another raspberry for mischaracterizing the security posture of the IBM i and System z mainframe, which have long been considered among the most secure platforms available to businesses. The authors insinuate that “code-based vulnerabilities” are lurking in the operating system, waiting to wreak havoc on innocent users.
“Although there are solutions that can scan for these code vulnerabilities, they are not capable of finding them in the larger operating system,” they wrote. “OS-level vulnerabilities are a bigger problem, due to potential hacker exploitation that gives access to data, applications, and users across the mainframe. A single flaw can result in negative effects across hundreds of applications and thousands of users.”
The authors are right that a security vulnerability in a mainframe or an IBM i server could have devastating consequences. But that’s because organizations still run so many of their applications on these workhorses to begin with, not due to some hidden security problem within the system software.
There are real challenges facing the IBM i and System z platforms, and the companies that rely on these computers to run their businesses. Keeping a solid pool of skilled technicians available, maintaining a modern and integrate-able code base, and ensuring proper security configurations are all top-of-mind for CIOs at companies that rely on IBM i and mainframes to run their businesses.
What FNTS failed to recognize in its e-book is that these are universal concerns that impact CIOs for all organizations, no matter what platforms they use. Migrating to an X64 architecture may be in the best interest of a given company looking to buy or build cutting-edge applications to drive their business forward, but it won’t eliminate these concerns from the mind of the CIO.
FNTS is correct in that many IBM i and mainframe shops are struggling to chart a course forward toward a modern computing architecture. However, in many respects, the problems are largely due to a lack of investment in IT on the part of the customer that has spanned for many years, if not decades.
Unfortunately, too many IT executives have mistaken the well-understood reliability and longevity of IBM i servers and mainframes as a license to do nothing, or to do less anyway. As a result, the applications that run on the platforms have grown old, difficult to maintain, and largely incompatible with modern computing standards.
IBM has done a relatively good job of giving IBM i users the tools and technologies they need to create modern business applications that can hold their own with any platform from an aesthetic and functionality perspective, while still maintaining its long-held advantages in stability, maintenance costs, and security. However, Big Blue has done an even better job of allowing customers to continue to run 20- and 30-year old application code, year in and year out.
Don’t blame IBM for the extreme level of backwards compatibility that it’s given users. After all, it’s just giving customers what they want. By overlooking the positives of the modern IBM i server and instead focusing on servers that have been neglected by their users, FNTS does the midrange and mainframe communities a disservice.
FNTS did not respond to emails and telephone calls seeking answers to IT Jungle‘s questions about the e-book. You can download a copy of the e-book here.
Modernizing The IBM i Estate For Digital Transformation
IBM i’s TCO Advantage Widens, According to Reports
Modernization or Migration? Survey Aims to Sort Out the Direction
IBM Redpaper Guides You To IBM i Modernization Tools
Modernization? How about NEW? There’s been nothing new in QSYS since ILE.
I can’t think of a vendor developing any NEW business applications (not modernized existing business apps, not technical tools) in QSYS. Who can blame them? There is no coherent UI direction from IBM. Other than 5250 there is no UI *guaranteed* to be available. It’s great to support various UI options but there must be a graphical UI that vendors know is available on every machine, otherwise there’s compatibility issues for customers having software from more than one vendor.
It’s great that software can be integrated via PASE. Nobody is going to buy an IBM/i to run apps in PASE only. People who buy IBM/i already have apps in QSYS. And there’s nothing new in QSYS. And the ecosystem shrinks.
Having a graphical UI guaranteed to be available will be a tremendous boost for the machine. Without it the machine is dying. It’s a superior machine, but this deficiency is killing it slowly but surely. Spare me any cheerleading, I like i, but this problem has got to be fixed. Then maybe there will be NEW apps running in the glorious QSYS. That means new vendors, new customers, and a growing ecosystem. For IBM/i to thrive there has got to be new development in QSYS. Nobody is going to buy an IBM/i just to run *nix apps ported from far less expensive hardware.
New apps in QSYS is future growth, but there’s got to be a GUI. Without it we die.
QSYS or bust!
How not to agree to this post?
This just hit the biggest point of all. It’s not too late, give a basic for business use UI, standard and supported by IBM, and well integrated with RPG and the programming model and transactional semantics, nothing fancy, say SAPGui-like in scope of visual tools. Integrate it in the classic ACS client besides the 5250 terminals and tools.
Will be really a great platform to develop for, really a turnkey solution with almost what it is needed for a business.
I’m really tired being presented yet another “modernization” tool to translate displays to html or – worse – couple or depend the code to, taking also in account the different semantics in protocol and transaction management in 5250 and http.
Without a standard GUI integration in IBM i , IBM will not attract new development and surely will slowly and slowly bleed…
Thanks Alex I think this a really solid review and its important to call out the obvious raspberries. Although I cant comment for the IBMi community it is worth noting that there a number of ISV’s and others that are enabling System z to integrate with “the rest of the systems”. The approaches are much more sophisticated than the screen scraping that has been prevalent in the past. The ‘Open Mainframe Project’ being a recent case in point. For our part are also exposing various mainframe subs-systems as RESTful and are working hard to break down perceptions that the mainframe remains a barrier to digital transformation.