Clearing Up IBMers’ Biggest Misconceptions About AWS (and Vice Versa)
James Roarty, chief cloud officer at Lightedge, has seen things from both sides, explaining how he went from being 'all in on cloud' to a more nuanced view
“They are two very different worlds.”
This is how James Roarty, chief cloud officer at Lightedge, began a conversation with TechChannel about the misconceptions IBM Z and Power professionals often have regarding AWS. That conversation, however, went both ways, since AWS pros have their own misconceptions about IBM environments.
Over the course of his career, Roarty says he has gone from being “all in on the cloud” to “hold on, maybe not.” He previously worked for Connectria—where his specialty was “all things hyperscaling”—before it was acquired by Lightedge, an AWS Premier Partner and one of the top three IBM hosting partners globally.
Lightedge provides hybrid enterprise infrastructure, so over the past two and a half years as chief cloud officer, Roarty has become familiar with the “IBM side of the house,” giving him an unusual understanding of both platforms and the misconceptions held by professionals specializing in either ecosystem.
Roarty and TechChannel discussed some of the most common and widely held misconceptions on both sides of the question.
The Lift-and-Shift Myth
AWS professionals think you can just “move everything over,” but it’s not so simple.
“My first experience in the infrastructure cloud world was very much ‘everything goes to public cloud,’” Roarty says. Now it’s important to understand that “going all in on any one platform may not be the right plan,” he adds.
It’s entirely possible to move VMware workloads to AWS via a lift-and-shift methodology. However, it can still be a big endeavor, especially if you’re moving thousands of VMs, even if you’re not rewriting any applications or redoing any architecture.
Roarty explains:
“You can’t just drag and drop IBM resources into the hyperscaler the way you can many x86 workloads; it’s a different underlying technology. Moving them natively to AWS or Azure would typically require significant application modernization or rewriting. That is a multi-year endeavor.”
But when it comes to IBM and the Z series, the i series or AIX, there’s not a methodology for lifting and shifting to hyperscalers or to AWS. The problem becomes even more complex when a business’ core systems—such as inventory tracking, revenue management or ERP systems—are running on those IBM platforms. In those situations, it could cost millions of dollars and take years to move everything. The most important question to ask in such circumstances is, “Why? What’s to be gained in making that move?”
AWS proponents may not realize the advantages of sticking with what works.
If a mission-critical workload is stable, reliable and meeting the business need, rewriting it simply to change where it runs may not create enough business value to justify the cost and risk. When the same platform performs a particular workload every night for 40 years and has never broken, what’s the advantage of moving it? What’s to be gained from taking that risk?
“If a company were to come to me and say, ‘We have these servers that do the exact same thing and run the same job every night, and we just need it to do this thing,’ I’d tell them it’s probably cheaper to keep it in a private cloud environment where it’s stable and running,” Roarty says.
The Data Advantage IBM Shops Miss
IBM supporters sometimes see the hyperscalers as just a different data center, but that view misses some of the biggest advantages of using AWS.
They may ask: Why store your data on someone else’s property when it’s fine on-prem? This misconception can be costly, since AWS offers more than storage.
“You have all this untapped data just sitting there,” explains Roarty. He offers a scenario: A company has seven orders delayed. With AWS capabilities, you can begin to do queries about why those orders are delayed. Are there specific points in the month where the delays occur? He says that some teams have built tooling internally to do some queries, but “there are a lot more services natively available from AWS, out of the box.”
Companies have collected and stored data for years, and now the tools are available to gain actionable insights from that data. The cloud-native tools that support the hyperscaler environment can yield surprising information and reveal unexpected patterns. Simply asking, “What are the anomalies in this data set?” using tools like Bedrock from AWS can result in crucial information that may have remained unknown.
IBM professionals often assume on-prem or a private cloud is more secure.
Roarty observes that one of the stickiest and most widely held misconceptions among IBM professionals is that data stored on-prem or in a private cloud environment is more secure than data in a hyperscaler. He points out two problems with this belief.
First, many governments and governmental agencies, as well as financial institutions, healthcare organizations and others operating in heavily regulated industries, use AWS. At this point, “the security’s there,” Roarty says. He also notes that AWS holds numerous certifications attesting to their security across numerous regulated sectors.
Second, he notes that an on-prem environment can be poorly architected, making the data more vulnerable than it would be using AWS. Simply being on-prem doesn’t secure the data. If you have a solid door on your house but you never lock it, your possessions aren’t more secure than they would be in a storage facility with a gate and automatic locks.
Two Things Everyone Gets Wrong
Having all of your eggs in one basket is, generally speaking, a bad idea.
Roarty notes that everyone tends to overlook a couple of important ideas.
At one time, it may have made sense for an organization to be dedicated to a single platform, but that time has passed. Today, it makes much more sense to spread things out so that each workload is where it delivers the best combination of performance, resilience, economics and innovation. Operations, storage, analysis and workloads should be done where it makes the most sense.
When cloud-based AI tools can provide actionable insights that would otherwise remain obscured, it may make good business sense to use those tools. When a business-critical workload has been running reliably on-prem, keep doing it that way.
“I don’t think there’s a one-stop shop for anything,” Roarty says. The appeal of a single vendor is clear, but it’s not the most efficient or logical way to approach business in the modern and complex world.
“Modernization” can mean different things, and it’s crucial to understand the implications of any modernization plan.
Misconceptions about platforms are plentiful, which can make it difficult to discuss modernization. It’s even more difficult when it becomes clear that the very definition of modernization differs—sometimes dramatically. Roarty once thought modernization meant moving everything to the cloud, but that opinion has changed with time and experience.
“Putting your eggs in one basket,” Roarty says, “is not the best way to run tech for an enterprise organization.”
The divide between AWS pros and IBM advocates is real, but not unbridgeable. Focusing on the strengths of each and making decisions from a business-first perspective limits risk and may reveal unexpected opportunities neither side could have predicted. As Roarty suggests, “Put resources where they make the most sense.”