Digital Next Advisors and HPE Pointnext Services Advisory and Professional Services experts map technology and offer best-practice advice to help you: deliver on digital ambition, create new value streams, speed up innovation, improve customer experience, and more.
The Why, What, How, and Where of Offloading from Mainframe
BrandPosts are written and edited by members of our sponsor community. BrandPosts create an opportunity for an individual sponsor to provide insight and commentary from their point-of-view directly to our audience. The editorial team does not participate in the writing or editing of BrandPosts.
By Ian Jagger
AWS, Microsoft, and Google each have mainframe migration programs. The cloud versus mainframe debate has been running for years, but there’s a reason why cloud service providers are nibbling at the mainframe user market, well actually, several reasons.
The global compute market is expected to reach $505.45 billion in 2025 at a CAGR of 8% and 2021 is estimated to have a CAGR of 10.9% over 2020. In comparison, the mainframe market has a CAGR about half of that.¹ This translates to movement away from mainframe to Linux-based systems.
Mainframe users tend to be incumbents with data-heavy and highly critical applications (think banking). Innovation in cloud is faster, isn’t it? How do these incumbents fend off disruptors, for example, credit card companies’ market share versus the significant growth in mobile wallets, shopping cart instalment apps, and electronic payment applications?
Companies placing customer experience strategies first outperform laggards (those placing other strategies first, such as cost) by 80%. Further, coming out of 2020, two-thirds of CEOs said they are “more confident” in the growth trajectories of their companies over the next three years than they were at the beginning of the year.³ Therefore, speed of innovation is critical.
There is an increasing shortage of mainframe-skilled workers, while for mainframe development skills and customer-specific application knowledge skills the gap is increasing. The same digital demands around customer experience are driving unrelenting demands on mainframe development. It’s estimated only 37% of outgoing mainframe talent has been replaced over the last five years.²
As mainframe utilization lessens, the costs associated with running mainframe systems increases, in an already “criticality, but at a huge operating cost” world.
Targeting from cloud providers is often driven by their particular flavour of cloud being expanded, with solutions like AWS Outposts, Google’s Anthos, and Azure Stack that bring the cloud closer to the mainframe rather than the other way around. Organizations are looking to leverage hybrid cloud models like SaaS/PaaS that can enable to scale cost-effectively, with improved application development capabilities. The future will favor solutions that facilitate the adoption of these hybrid models. For this reason, AWS has partnered with Micro Focus and Google acquired Cornerstone. It’s hard for IBM to argue the case given that by the end of 2021, IBM will spin off a sizable portion of its managed infrastructure services business to focus on cloud computing.4
It’s a tough call for mainframe users. A lot of businesses still rely heavily on their mainframes to power mission-critical workloads. Moving them to alternate platforms without a proper investment of funds, skillsets, and time can be a costly – and a risky process if done incorrectly. Does this mean mainframe users are stuck with it and won’t be able to realize the benefits of a “cloud everywhere” strategy?
Why offload from mainframe
Here are the key reasons to make the move – you might have more!
Digital transformation: More than 85% of Fortune 500 CEOs are planning to or have already begun to accelerate their digital transformation programs after the experiences of 2020/21. Two core reasons are the disruption already experienced, but also the recognized opportunities provide the vision, and it is the digital disruption of new customer experiences that is bolstering revenue streams. For this, agility is required.
Operating costs: The hardware, operating system, database, and third-party application costs associated with mainframes are high, and they continue to increase.
Footprint and power costs: Standards-based hybrid cloud platforms now match the performance of mainframes while using less power and having a smaller footprint.
Availability of skills: Staffing needs are much easier to meet when you run workloads on hybrid cloud platforms. As the mainframe operations workforce gets older, these skills may no longer be available.
Inflexibility: Mainframes have traditionally been run in a hosted IT environment—a structure that makes it very expensive to scale.
Choice of applications: Thousands more up-to-date applications are available on Intel-based servers than on mainframes, offering greater choice to meet IT and business needs.
Utilization: Hybrid platforms utilize a range of virtualization tools that enable you to enhance consolidation of workloads and improve utilization levels.
What needs to move?
No two organizations are the same, each has its own IT landscape and degree and maturity of cloud adoption. The degree and extent of migration benefits can vary from one organization to another. If you are a mainframe user, there would be major nervousness at undertaking a comprehensive lift and shift. A CIO using mainframe would argue certain workloads are better suited to each system. Firstly, hey should frankly ask, “what is in my mainframe?” There’s an honesty about recognizing when the last deep assessment of the application library took place. Perhaps Y2K? Mainframe migration should start with looking at what you currently have, understanding the relationships and inter-dependencies.
The identification of the right migration strategy is crucial for the success of modernization. A strategy begins with discovery. What do you have? What is your outlook? Do you have a target destination? After these assessments have been made, start developing the stepping stones to get you to the place you want to go – while mapping it, implementing, and deploying it. A method to think about this is to break it down:
Business discovery: top-line catalogue of what IT services are required for the business
Application discovery: application functions and detailed components to provide the IT services
Influences discovery: detailed information on all elements that provide each component and function
Then consider them in two groupings:
“Need to work” – Run the current business, day-to-day business needs
“Need to win” – Innovate and create competitive advantage
Armed with this information, you can determine how to prioritize workload for migration.
How to undertake the migration
Bear in mind that everything on a mainframe has an equivalency elsewhere. For example, on the IBM mainframe, you have the Independent Service Processor, in HPE GreenLake, you have GreenLake Central. If your migration strategy is correctly planned, anything can be migrated. It’s a question of addressing the notion of the “ease and impact” of moving any given application when measured against criticality, interdependencies, and a host of other measures.
Moving off mainframe requires deep and accurate understanding. This article began with the concerted attack on mainframe environments from cloud service providers. But they own the environment, not the “how to” expertise. What is astonishing is the recent pre-announcement from AWS regarding the launch of the AWS Mainframe Migration category within their Migration Competency. They are claiming this migration expertise through their partners “with mature practices and a track record of successful mainframe migrations.”
Let’s take a closer look at that “expertise”, Infosys, Deloitte, and Tata are listed as “consulting” partners, which is fair, but they are not the companies being placed as qualifying for the AWS Mainframe Migration Technology category. AWS pre-announced four partner companies. A quick LinkedIn search tells me that one of them has 38 employees, one was a UK-focused software company until a US acquisition two years ago, a third has 28 employees, and the fourth is Micro Focus – a $9 billion merge-spin from HPE and its pedigree.
Offloading from mainframe requires deep and proven experience. I’m biased, but in comparison, HPE has been doing it for 25 years. Its Mainframe Migration Factory Center of Expertise uses proven migration methodology, automated tools, and frameworks for solution design and migration execution. The global team is approaching 500 platform migration projects, mostly of them critical applications, including migrations of up to 85,000 MIPS. And, for those looking to migrate to reduce cost, which is by far the primary driver, the average achieved by that migration base is 70% reduction in TCO, with an ROI averaging less than 24 months.
Where to migrate
Mainframes are known for workloads with a lot of data gravity, latency, application dependency, regulatory compliance, and they lack the agility of the modern cloud experience. Now you can bring cloud speed, agility, and an as-a-service model to where your apps and data are located. Today, organizations can operate with one model and one experience across distributed clouds, for apps and data at the edge, in colocations and their own data centers.
While the hybrid model seems to be the way forward, there are methodologies around how to follow a “cloud everywhere” model. As we have seen, allocating tasks to each platform relies on an accurate assessment of the workload of a particular business. This approach is more effective that an “all or nothing” embrace of either cloud or mainframe systems and relieves the CIO of wrestling with that decision.
However, today, this is just the beginning of the journey – it’s also about going beyond that and being able to consume advanced services.
Mainframe utilization is no different from any other platform. Organizations don’t run at 100% utilization every day, all day. With HPE GreenLake, and bear in mind, this is in any cloud environment, on or off-premises, you only pay for what you use. By only using additional capacity when you need it, you don’t have to keep that capacity in reserve and pay for it when you don’t need it.
It goes beyond that. How about unlocking data value? Organizations can leverage the HPE GreenLake framework, for example, to expose dormant data and move it into into AI and machine learning to see trends and react to what you now know. When we think about turning data into action, it really is about new functionality and new capabilities to do machine learning and AI.
Are cloud service providers wrong?
Are cloud service providers wrong to attract mainframe users? Absolutely not, there are valid reasons to draw users to the “cloud-side.” Mainframe users by now must be querying what the advantages are, but I would suggest that, with their minds open, and to quote the oft incorrectly-quoted proverb, they can eat their cake and have it – it just happens to be colored green.
See the“Mainframe to HPE GreenLake” ” infographic here.
¹Absolute Reports: Global Mainframe Professional Survey Report 2021, Forecast to 2026
² For Mainframe-Reliant Organizations, A Perfect Storm Is Coming. Forbes Technology Council April 2020
³ KPMG Survey August 2020
4wikipedia.org/wiki/List_of_mergers_and_acquisitions_by_IBM, 5www2.deloitte.com insights technology focus areas to accelerate digital transformation.
About Ian Jagger
Jagger is the creator and writer of HPE Pointnext Services narrative, focused on digital transformation, linking technology capabilities expertise with business goals.
A Chartered Marketer, his experience spans strategic development and planning for Start-ups through to content creation, thought leadership, AR/PR, campaign program building and execution for Enterprise. Successful solution launches include HPE Digital Next Advisory, HPE Right Mix Advisor, and HPE Micro Datacenter.