Is The Cloud On Your IBM i Horizon?
September 20, 2023 Alex Woodie
It’s almost impossible to ignore the enormous pull that the cloud is exerting on the software industry today. Companies that run the IBM i server are also subject to that gravity, although it has different ramifications for their technological future, as well as the decisions they’ll make if they want to run their operations in the cloud.
IBM i shops, almost by definition, are not running on the cloud today, at least not in the public cloud the way it is normally defined. While there are a handful of managed service providers (MSPs) offering private cloud IBM i runtimes, and some SaaS vendors delivering IT services atop IBM i servers, IBM i shops do not have the option to “lift and shift” their homegrown IT and application stack into the public cloud providers, AWS, Microsoft Azure, and Google Cloud, the way that companies running on industry standard X86 servers can. (IBM Cloud does provide an IBM i runtime, but it lacks the humongous infrastructure and economics of scale that the big three offer.)
Of course, that doesn’t mean that the cloud is not in IBM i shop’s futures. It almost certainly will be. The difference is that IBM i shops almost certainly won’t be bringing their existing IT and applications stack with them, at least as they currently exist.
The massive growth of the public cloud and software as a service (SaaS) industry over the past 10 years is impacting all of IT. According to a recent Gartner report, the public cloud will see nearly $600 billion in spending this year across SaaS, platform as a service (PaaS), infrastructure as a service (IaaS), and other as-a-service (*aaS) offerings. That’s up from nearly $500 billion last year, representing about a 21 percent growth rate. Next year, Gartner estimates that a 21 percent growth rate to hold steady, yielding $725 billion in cloud and *aaS spending.
The SaaS market has exploded over the past decade years. In 2015, the world spent just $31.4 billion on SaaS offerings. This year, that number is projected to be $196.3 billion, representing a 5.2x increase in spending over nine years. As you can see from the Statista graphic, the curve is bending upward, indicating an acceleration of SaaS spending.
Clearly, the SaaS market is a hot one at the moment, as traditional software vendors and startups move to offer their software on the cloud. By some accounts, there are over 70,000 SaaS vendors operating in the cloud worldwide, although the counts in the 25,000-to-30,000 range seem to be more frequent.
What does this explosion of the public cloud and SaaS mean for IBM i shops? Well, there are good aspects and not-so-good aspects to it.
On the plus side, the availability of advanced analytics and AI solutions in the cloud has never been greater. Companies that move some of their data into a public cloud have a plethora of offerings available to them. There’s no need to build out an expensive on-premise data warehouse or GPU cluster to crunch data or build machine learning models anymore, as these systems are openly available in the cloud. All that’s required is hooking up some data pipelines from the production IBM i database and sending the data to a dedicated analytics repository, which is something that IBM i shops have been doing since the dawn of time (the IBM i’s underutilized and underappreciated capability to serve as a data warehousing machine notwithstanding).
Where the negatives start to creep in is mostly on the operational or transactional side, which has always been the IBM i’s sweet spot. The midrange server and its integrated database have traditionally served as the backbone for homegrown and vendor-supplied enterprise applications that support the day-to-day operations of businesses in all sorts of industries. While business-critical production systems have been slower to move to public clouds than other workloads (such as analytics, app dev, testing, backup and recovery, and high availability-disaster recovery), there’s no denying that there is a slow and steady momentum to get them into the cloud.
There are a couple of issues for IBM i shops considering moving production business apps and ERP systems to the public cloud. If it’s a custom IBM i application, you’re either going to be running it in a private cloud or rewriting it to run on the x86 systems in the public cloud. And if it’s a packaged application, you’re almost certainly going to be experiencing a reimplementation of the product, since IBM i doesn’t run in the cloud.
Miten Marfartia, the CEO of modernization tool provider EvolveWare, says there are solid paths for IBM i shops to rewrite their existing apps to be cloud native.
“When it comes to modernizing legacy applications and taking them to the cloud, the best approach is the cloud native approach,” Marfatia tells IT Jungle. “There are two options: (1) Extract business rules from the source application and either implement them in a cloud enabled commercial off-the-shelf (COTS) product, or re-write the application using state-of-the-art technologies, generating microservices in the process. (2) Use an application modernization platform that can generate modern code including microservices using current technologies.”
Rewriting monolithic legacy applications to use microservices is particularly important due to the architectural differences between the public cloud and the IBM i or System Z systems they originated from, Marfatia says.
“A large number of legacy applications that require modernization for the cloud are monolithic batch applications processing huge volumes of data in unattended mode, i.e., the code processing data is single-threaded or all in one piece as a single copy,” he says. “These applications are able to process such large amounts of data in the time period they do, because of the hardware power of the mainframe or midrange systems they run on. In order for the modernized application to perform similarly, the approach should be to separate the code that performs processing into microservices that can be installed as multiple instances or as multiple copies. In other words, you have multiple copies of the code processing data in parallel.”
IBM i shops that run packaged ERP applications from independent software vendors (ISVs) have a different set of variables to deal with. Many of these ISVs have committed to running their software on the IBM i operating system for years into the future – and in some cases, for decades.
Jack Henry & Associates, for example, has developed core banking applications to run on the IBM i server for decades. The platform has worked for the company and its clients very well, and Jack Henry has even served as a private cloud for many of its IBM i clients, managing the Power Systems hardware that their banking customers run on. But the company has kicked off a modernization effort that will see the company eventually move away from native IBM i development in favor of running software as containerized microservices in the public cloud.
Jack Henry customers likely have decades to prepare for any migration away from the IBM i server, and there’s a lot that could happen in the meantime. Maybe by 2040, IBM will figure out how to make Kubernetes work on the IBM i platform? Or maybe it will fit the platform into a container managed by K8S? There’s a lot of time to figure that out.
But not all IBM i shops have the luxury of lots of time. SAP, for example, is moving ahead with a plan to migrate all of its ERP customers to its new cloud-enabled platform, S/4HANA. The German software giant hasn’t budged on its 2027 deadline to end mainstream support for ECC and Business Suite, which is relied upon by thousands of customers running a variety of platforms, including about 1,500 or so on IBM i. Customers will be able to purchase extended support contracts through 2030.
The challenge for customers facing this deadline isn’t just the forced march to S/4HANA; it’s the fact that S/4HANA doesn’t offer the breadth and depth of capabilites they currently enjoy in their existing products. SAP has promised that the functionality gap is narrowing and will be gone by 2027, but some customers aren’t so sure.
Gartner says that by 2026, 75 percent of organizations will adopt a digital transformation model predicated on cloud as the fundamental underlying platform. That may be true, but it doesn’t necessarily mean that the digital transformation will be easy or without speed bumps.
Eric Kimberling, the CEO of Third Stage Consulting, has seen plenty of ERP implementations and digital transformation projects go haywire over the years. In his new book, “The Final Countdown: Strategies to Reach the Third Stage of Digital Transformation,” Kimberling recounts the story of a large American steel manufacturer that underwent a major digital transformation and ERP implementation project.
In the middle of their rollout, the ERP vendor transitioned their flagship product to the cloud. “Despite the software vendor’s attempts to persuade them to switch to the cloud solution, the client conducted a thorough evaluation and determined that the cloud version was not mature enough and did not meet their needs,” Kimberling writes. “They opted to stick with the older, more mature version of the software, which proved to be a better fit for their requirements.”