Updated: Nov 28, 2022
The Headlines read … “Mainframes will be Replaced” and “Everything is Moving to the Cloud” ...
For many years, companies have been reading these headlines and investing in digital transformations with hopes to modernize applications and eliminate costly mainframe filled data centers by moving to the cloud. Many companies in the past announced they will be “100% off mainframes in 3-5 years” (I actually started seeing these announcements almost 20 years ago). As the headlines say - “everything is moving to the cloud, so be ready and act quickly or you and your company will be out of business!”
Well, although we are still hearing and reading these headlines, in 2022 the data shows that everything didn't move to the cloud - especially in a few large industries like finance, healthcare and government insurance (there are more). Why not? On the surface, it seems like a no-brainer that eliminating a data center with many millions of dollars of gear and high operating costs, and being able to modernize applications by moving to the cloud where services are easily managed and accessed just makes financial and business sense. Also some companies in less mainframe dependent industries have done it successfully - so can we right? Not so fast ... So what actually happened with mainframes and what’s next? Will we ever get 100% of mainframes into the cloud?
What Happened so Far?
With the headlines saying everything will be in the cloud - The mainframe was clearly right in the line of sight and targeted for significant workload migrations and re-platforming. Sort of a “sitting duck” that was and is right there for us to gain incredible value and savings right? Well not exactly. Like most technology transformations, it’s never that clear - especially for mainframes that handle large amounts of processing - which is why we see 70% of these types of digital transformations fail to meet expectations. Don't get me wrong, there are successes, but the more complex systems often fail to meet full expectations ... if they try at all.
The reasons for this high failure rate are numerous; some of these are the same reasons many companies don't even try or give up quickly. Here are a few:
Don't fully understanding the complexities and performance of the mainframe and how much the cloud can really handle
Tightly coupled code not easily migrated
Skill sets and old code that programmers don't know these days
Operating system dependencies that don't exist in the Cloud
Legal, vendors contracts and lock-ins
Monolithic devices and systems
And the list goes on … often realizing what they thought would be an easy migration and a quick hit was really a major effort and would require a costly refactoring - if it can be done at all! To top it off, in many of these instances, the workloads targeted for migration went through the process only to have to fall back to the mainframe which also required work to put in back. In some cases the workloads promised new capabilities so the fallback required additional processing creating new demand for additional mainframe capacity. Hence a failed and costly digital transformation transpired with new problems to boot.
These failures not only cost money but also lost credibility with customers and leadership creating future problems for many firms. The next digital transformation - even if planned better - will be harder to fund, get corporate wide buy in and get started.
One Way to do it Better
I have been involved with a few digital transformations including those where mainframes were involved over the last few decades. In general, when it comes to any legacy technology transformation to modern technology - do NOT listen to the headlines or hype - yes sometimes it’s hard to see when something is hype or not, but you have to make sure your focused on driving a technology strategy that is delivers on specific business needs and not what vendors are headlining.
Cloud vendors want to sell cloud services and companies like IBM want to sell mainframes and so on - that’s their job - so you cant blame them! Your job as an IT or business leader, as you know, is to drive value for your customers by ensuring you are delivering a strategy that enables business goals. But the content of what you do and how you do it will define your success as in most things in life. In this case, reading the headlines that “mainframes will be replaced”, or “everything is going to the cloud” or the “data center will go away” is really irrelevant to what you need to do for your company. Some of the important questions you should ask yourself are:
How to use new technology in the future?
Is there new technology that can help achieve business goals better and faster?
What is the technology strategy and does it align to business goals?
What vendors own the future?
Who is the best partner to help us along this journey?
Collect and analyze internal and external data for the insight you need to make the right business decisions. identify the problems you have to solve, develop and prioritize actions, align to your business goals, develop metrics, and get actions funded
and moving forward focused on enabling business goals. If you do this right, you will be focused on your business and if the headlines are correct - it will be part of your transformation anyway.
My point isn't to ignore these headlines, it's really to understand they are less relevant than what the people who put them out there want you to believe. As a business or technology leader, you need to keep these headlines in the back of your mind and go through the proper data-driven process outlined above to identify the actions you need to deliver what's best for your business.
Very often you will find the answers are a combination of things and usually not 100% in one direction - for example, when the cloud vendors first headlined “everything will go to the cloud”, they were focused on public cloud services. But as companies failed to meet expectations and redeveloped new actions that were working better, it was often a hybrid of private (really on-premise) and public cloud developments that were working. Now it's evolving again to even newer configurations - also a blend of things. So real answers come from proper technology planning as well as understanding the static one sided view of these headlines are really irrelevant.
Now this is easier said than done for many. Headlines like these are powerful and become mainstream pretty quickly and usually get a lot of executive support. I think besides developing a technology strategy that incorporates people, process, tools and technology, you also need to adopt the right mindset as it pertains to these headlines around mainframes and cloud (data center). This will help you articulate to executives, boards and stakeholders what needs to get done to be successful versus the hype:
Adopt the right Mindset
Perhaps the future goal of mainframes in the long term is that they will indeed be replaced. While this may be true in the long term (and that's debatable in itself), you probably don't want to focus on that being the first or even second goal in the short and medium term. If you're like most companies, your mainframe is running your organization’s most critical business applications and perhaps handling a very large volume of transaction processing workloads. Your goal should be focused on improving and advancing your system to achieve business outcomes - and this is true for any technology effort.
You really have 2 choices on how you want to lead this effort. The mainframe mindset you choose will determine the outcome:
Eliminate the mainframe and move everything to the cloud focused on cost
Deliver newer and better versions of the system focused on business outcomes
Well a third is do nothing - but we wont go there.
Which one should you choose? I think you guessed it …
To be successful with a migration of mainframes, you need a strategic plan which defines an evolution of new versions of the system versus a one time big bang effort. This should include new versions of the mainframe and its applications essentially forming a piece of your hybrid cloud evolution versus it being a big bang effort to to replace your mainframe.
Adopting an evolutionary mindset when it comes to how you modernize mainframes and the associated critical applications into their next, newer and better versions is necessary in order to be successful. So you need to change your mindset from a goal of “replacing mainframes” focused on cost to adopting a mindset of “evolving mainframes to newer and better versions” focused on achieving business goals and value. This may result in the mainframe eventually being completely replaced, or it will become an important part of your hybrid cloud. Either way, it will do so based on this evolution to newer versions and business outcomes and not based on hype or unattainable big bang goals.
Data Center (Cloud) Mindset
When you think of a cloud migration of any type, you think of the cloud being an “on-ramp” for your legacy applications to run on whatever they need to run on to achieve business goals. The on-ramp decision process takes into account all business needs and requirements including reducing costs, capital depreciation and IT operations inertia, data security, new features and so on. So, the evolution of application workloads “on-ramp” to the cloud is the focus as determined by your mainframe mindset.
So what about the data center? Aren't we supposed to eliminate that too? I believe you should not have a specific project plan to eliminate the data center. This will cause a conflict between implementing newer and better versions of your applications with the goals of eliminating a data center. The two will quickly disconnect and you will become another statistic for the 70% failure bucket. The evolution of a data center will happen by the newer versions determined by the migration. The data center will either be replaced, or become obsolete, or become a valued member of your hybrid cloud - but you will see this happen as defined by the cloud migration per your technology strategy - these things stay coupled versus being implemented separately.
So the data center mindset? There isn't any, your mainframe mindset incorporates data center evolution as defined in your technology strategy.
Evaluate Relationships with a Managed Infrastructure Provider (Kyndryl)
For those of you that may not be aware, Kyndryl was a division of IBM until November 2021 when it was spun off into a separate company.
Today, after some bumps in the road, Kyndryl is considered by many as one of the largest IT infrastructure providers in the world. This is due to the fact that Kyndral is becoming much more independent of IBM developing new services and partnerships with large cloud providers; earlier this year, they signed a deal with AWS for cloud enablement providing the tools and cloud migration programs to help clients successfully migrate mainframe workloads to the cloud. They have similar partnerships with the other major cloud providers as well (Microsoft & Google). There have been success stories already documented with mainframe migrations.
I'm not saying run out and forge a relationship with Kyndryl without following proper evaluation within your company standards - I am saying you should take a look at Kyndral as they do come from a mainframe company and support a very well established IBM mainframe as a service offering along with cloud tools, partnerships and so on. This will become a real enabler moving forward.
You may want to consider bringing in a vendor agnostic partner to help you in this area as they can dig deep into what is needed to be successful. This could be the same person/company you use for this entire technology strategy and planning effort.
VNCAcademy will soon release an online event to discuss the pros and cons of a Kyndryl type relationship. The event will include discussions, articles and debates with the real experts. You need to be signed up to get FREE access.
Technology Strategy (Digital Transformation)
With your mainframe mindset well established as noted above, you're ready for defining your technology strategy to enable the digital transformation. This is basically about defining a technology strategy aligned to business goals resulting in an actionable digital transformation strategic plan. If you haven’t done this before, I highly suggest partnering with a technology strategy advisor and implementation firm/expert to avoid very costly mistakes. The creation, development and implementation of this plan hinges on having the necessary skill sets, experience and leadership.
As your digital transformation plan will likely include migrating to the cloud, keep in mind that mainframe technology is evolving to handle AI and machine learning, IoT, edge systems, mobile applications, and cyber security threats. Legacy and monolithic mainframe systems that operate in isolation have evolved now integrating with mobile technology, web services and distributed applications running in the cloud. Remember your mindset is evolutionary which means implementing newer and better versions of the system and incorporating this new mainframe technology could help accomplish amazing, powerful, and new hybrid type opportunities for your clients. So right away, this moves from “replace the mainframe and move applications to the cloud” to “delivering business value in an evolutionary, measurable and manageable way”.
One important note on measuring success of your digital transformation to the cloud - it may be subtle, but put metrics in place that measures success in terms of new capabilities that were enabled that didn't exist before, Avoid the measuring success in terms of what gets moved to the cloud that shuts down the data center or mainframe. I see this mistake being made over and over again - it's subtle but often the results are negative because it tends to drive the wrong behavior within the company. This is important to help you maintain that evolutionary mindset!
As you go about your digital transformation with your mainframe evolutionary mindset, you will be enabling newer and better versions of the mainframe applications before you completely replace it (if you ever do - I believe this will be a hybrid for a while). You will get quick hits and start building up credibility not to mention achieving ROI. This may require upgrading your mainframe to a newer model to get some of the new features you may need - but the idea is to ensure you're delivering on your business goals versus trying to support the hype and replace the hardware and potentially not meeting business expectations.
Upgrade the Hardware?
This depends on your current hardware vendor, but since many of the mainframes are IBM - let's stick to this vendor. Although IBM has seen some slow years in mainframe sales, things may change for them in the near and medium term future. The new z16 IBM mainframe has massive processing advantages that you may want to consider depending on your transformation business goals. If your strategy is focusing on a hybrid model where the new mainframe technology fits in, you should evaluate this hardware as you do anything else in your business as it could be a differentiator.
Based on this article, you may believe I'm a fan of the mainframe or against moving everything to the cloud. This isn't true at all - I'm not against anything that makes sense - including major cloud migrations and sunsetting mainframes hardware - been there done that. Bottom line - I've had decades of success in my career around digital transformations because of my mindset and approach - and that's what I hope you are getting from this article. I never focus on hype, sound-bytes, or headlines - this stuff is irrelevant. I focus on my business and customers which drives me to use my experience and technology knowledge to create and relentlessly deliver a technology strategy that enables a digital transformation focused on business outcomes. It's really that simple folks. This delivers success for your company, customers and your own career.
VNCAcademy will shortly be releasing a series of podcasts that include discussions with executives and experts who have been through these sorts of migrations across multiple industries. These experts experienced both failure and success and will be debating and discussing what works and doesn't work. You have to sign up for this FREE event and to know when it's available.
So, to conclude, when I think about mainframes and the cloud, I don't think in terms of eliminating hardware or supporting the cloud vendors’ dreams of all of us magically moving everything to their cloud of choice. We should not define success based on what we are moving to the cloud for the sake of eliminating hardware or because we can't find developers for this old operating system and language. We also should not define success based on how much “private cloud” we implement in our data center. This is the wrong mindset which is why we see so much failure in this area!
We SHOULD think in terms of how vendors and technology can help my company achieve its goals. The evolutionary mindset brings you to another level and enables success short and long term. It avoids how you can help the vendor's bottom line, instead it allows you to think about how vendors and technology can help enable what you could not do before to deliver differentiation and amazing products to your customers. The result likely won’t be exactly what cloud vendors were hoping for in their headlines, but it will be a hybrid of something really great for your company, your career, and your customers.
Hope to see you all in VNCAcademy! Please join us - it's all FREE and really getting better and better every day.